Papers, code, etc. for the Deep Learning Study Group.
Meeting time - Tuesdays, 6:30 pm California time on Zoom
Zoom and Discord links are on the meetup page:
https://www.meetup.com/handsonprogrammingevents/
Weak-Driven Learning: How Weak Agents make Strong Agents Stronger
https://arxiv.org/abs/2602.08222
Recursive Language Models
https://arxiv.org/pdf/2512.24601
Blog
https://alexzhang13.github.io/blog/2025/rlm/
Github
https://github.com/alexzhang13/rlm
Documentation
https://alexzhang13.github.io/rlm/
ConceptMoE: Adaptive Token-to-Concept Compression for Implicit Compute Allocation
https://arxiv.org/pdf/2601.21420
Reinforcement Learning via Self-Distillation
https://arxiv.org/pdf/2601.20802
Conditional Memory via Scalable Lookup: A New Axis of Sparsity for Large Language Models
https://arxiv.org/pdf/2601.07372
mHC: Manifold-Constrained Hyper-Connections
https://arxiv.org/pdf/2512.24880v1
There are multiple YouTubes including:
https://www.youtube.com/watch?v=jYn_1PpRzxI
Background material: Hyper-Connections
https://arxiv.org/abs/2409.19606
Digital Red Queen: Adversarial Program Evolution in Core War with LLMs
https://arxiv.org/pdf/2601.03335
Website:
https://pub.sakana.ai/drq
Code:
https://github.com/SakanaAI/drq
There are many YouTubes on this work.
Hessian structure of neural networks
https://arxiv.org/abs/2505.02809
Blog: Loss functions and optimizers – Adam and Muon and the Hessian of the loss function
https://securemachinery.com/2025/12/18/loss-functions-and-optimizers/
When Models Manipulate Manifolds: The Geometry of a Counting Task
https://transformer-circuits.pub/2025/linebreaks/index.html
NVIDIA-Nemotron-3-White-Paper.pdf
https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-White-Paper.pdf
For addition background, if interested:
https://research.nvidia.com/labs/nemotron/files/NVIDIA-Nemotron-3-Nano-Technical-Report.pdf
The Path Not Taken: RLVR Provably Learns Off the Principals
https://arxiv.org/pdf/2511.08567
YouTube:
https://www.youtube.com/watch?v=iYpQJK5KLlw
Additional material
https://github.com/davidmacmillan/DeepLearningStudyGroup/blob/master/2025-12-23%20Supervised%20fine-tuning%20vs.%20reinforcement%20learning%20with%20verified%20rewards%20_%20Claude.pdf
1000 Layer Networks for Self-Supervised RL: Scaling Depth Can Enable New Goal-Reaching Capabilities
https://arxiv.org/abs/2503.14858
Additional background - Project site:
https://wang-kevin3290.github.io/scaling-crl/
Code:
https://github.com/wang-kevin3290/scaling-crl
Helpful CRL background info by one of the authors:
"Contrastive Learning as Goal-Conditioned Reinforcement Learning"
https://arxiv.org/pdf/2206.07568
PaTH Attention: Position Encoding via Accumulating Householder Transformations
https://arxiv.org/pdf/2505.16381
No meeting December 2 due to NeurIPS
Nested Learning: The Illusion of Deep Learning Architectures
https://abehrouz.github.io/files/NL.pdf
Blog on Nested Learning paper
https://research.google/blog/introducing-nested-learning-a-new-ml-paradigm-for-continual-learning/
DeepSeek-OCR: Contexts Optical Compression
https://arxiv.org/pdf/2510.18234
Kimi linear attention
https://arxiv.org/pdf/2510.26692
Slides: https://github.com/davidmacmillan/DeepLearningStudyGroup/blob/master/2025-11-11%20Kimi%20Linear%20%26%20Kimi%20Delta%20Attention.pdf
In-the-Flow Agentic System Optimization for Effective Planning and Tool Use
https://arxiv.org/pdf/2510.05592
Attention Sinks and Compression Valleys in LLMs are Two Sides of the Same Coin.
http://arxiv.org/abs/2510.06477
Less is More: Recursive Reasoning with Tiny Networks
https://arxiv.org/pdf/2510.04871
Bootstrapping Task Spaces for Self-Improvement
https://arxiv.org/pdf/2509.04575
Small Language Models are the Future of Agentic AI
https://arxiv.org/abs/2506.02153
Many YouTubes on this paper incuding by an author:
https://www.youtube.com/watch?v=9xgRTznP21E.
No paper this week. Instead we did an in-person social event (dinner) on Tuesday Sept. 30 at 6:30 PM in Mountain View, CA.
Real-Time Detection of Hallucinated Entities in Long-Form Generation
https://arxiv.org/pdf/2509.03531
Why Language Models Hallucinate
https://www.arxiv.org/abs/2509.04664
DataRater: Meta-Learned Dataset Curation
https://arxiv.org/pdf/2505.17895
A Survey on Diffusion Language Models
https://arxiv.org/pdf/2508.10875
GEPA: Reflective Prompt Evolution Can Outperform Reinforcement Learning
https://arxiv.org/pdf/2507.19457
Hierarchical Reasoning Models
https://arxiv.org/abs/2506.21734
There are multiple human YouTubes, including one by Gabriel Mongaras:
https://www.youtube.com/watch?v=TUsbk8vPDoM
Github:
https://github.com/sapientinc/HRM
Subliminal Learning: Language models transmit behavioral traits via hidden signals in data
https://arxiv.org/abs/2507.14805
Reasoning by Superposition: A Theoretical Perspective on Chain of Continuous Thought
https://arxiv.org/pdf/2505.12514
Mixture-of-Recursions: Learning Dynamic Recursive Depths for Adaptive Token-Level Computation
https://arxiv.org/pdf/2507.10524
Kimi k1.5: Scaling Reinforcement Learning with LLMs
https://arxiv.org/pdf/2501.12599
There are also multiple YouTubes.
Additional Kimi info, if interested:
Kimi-VL Technical Report
https://arxiv.org/pdf/2504.07491
DARS: Dynamic Action Re-Sampling to Enhance Coding Agent Performance by Adaptive Tree Traversal
https://arxiv.org/abs/2503.14269
Blog #1 - Gemma 3n model overview
https://ai.google.dev/gemma/docs/gemma-3n
Blog #2 - Introducing Gemma 3n: The developer guide
https://developers.googleblog.com/en/introducing-gemma-3n-developer-guide/
MatFormer: Nested Transformer for Elastic Inference
https://arxiv.org/pdf/2310.07707
There are multiple YouTubes on Gemma 3n and MatFormer.
MELODI: Exploring Memory Compression for Long Contexts (DeepMind, Oct. 2024)
https://arxiv.org/abs/2410.03156
Open Review:
https://openreview.net/forum?id=TvGPP8i18S
Relaxed Recursive Transformers: Effective Parameter Sharing with Layer-wise LoRA
https://arxiv.org/pdf/2410.20672
OpenReview:
https://openreview.net/forum?id=WwpYSOkkCt
Concise Reasoning via Reinforcement Learning
https://arxiv.org/pdf/2504.05185
Good news - no homework this week!!!
At the meeting, one of our members, Ted, will present MultiDecode,
original work he has done on speeding inference, including for RAG.
Efficient Sequence Transduction by Jointly Predicting Tokens and Durations
https://arxiv.org/abs/2304.06795
Fast Conformer with Linearly Scalable Attention for Efficient Speech Recognition
https://arxiv.org/abs/2305.05084
AlphaEvolve: A coding agent for scientific and algorithmic discovery
https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/AlphaEvolve.pdf
Blog:
https://deepmind.google/discover/blog/alphaevolve-a-gemini-powered-coding-agent-for-designing-advanced-algorithms/
Qwen3 Technical Report
https://github.com/QwenLM/Qwen3/blob/main/Qwen3_Technical_Report.pdf
There are many YouTubes.
Also try it out (e.g. Ollama has it) or here:
https://qwen3.app/
Flow matching for Generative Modeling
https://arxiv.org/abs/2210.02747
YouTube by Yannic Kilcher:
https://youtu.be/7NNxK3CqaDk
YouTube by Jia-Bin Huang (Univ. Maryland):
https://youtu.be/DDq_pIfHqLs
YouTube by Peter Abbeel (UC Berkeley):
https://www.youtube.com/watch?v=SkSDCzz41Vs
There are also other YouTubes and blogs such as:
https://www.youtube.com/watch?v=7cMzfkWFWhI
Round and Round We Go! What makes Rotary Positional Encodings useful?
https://arxiv.org/pdf/2410.06205
There is a YouTube from Gabriel Mongaras:
https://www.youtube.com/watch?v=2tS_bXPoriI
Why do LLMs attend to the first token?
https://arxiv.org/abs/2504.02732
As background, Evan Miller has a blog from 2023 on this issue and identified a simple fix:
add +1 in the transformer softmax denominators (but not to the final LLM output softmax).
https://www.evanmiller.org/attention-is-off-by-one.html
Tracing the heritage, tonight's paper on pg. 3 references Xiao 2024
https://arxiv.org/abs/2309.17453
and Xiao (pg. 4 & 6) notes his StreamingLLM approach for attention sinks can
(perhaps) be eliminated if one instead uses Miller's +1 softmax recommendation.
Yannic Kilcher has a YouTube on Xaio:
https://www.youtube.com/watch?v=409tNlaByds
Anthropic MCP
https://www.anthropic.com/news/model-context-protocol
MCP Introduction, Tutorials, Concepts:
https://modelcontextprotocol.io/introduction
Google agent2agent
https://developers.googleblog.com/en/a2a-a-new-era-of-agent-interoperability/
A2A Technical Documentation:
https://google.github.io/A2A/#/documentation
A2A and MCP:
https://google.github.io/A2A/#/topics/a2a_and_mcp
We are continuing the discussion from last week on the recent Anthropic papers/blogs.
We are doing the second paper/blog this week:
On the Biology of a Large Language Model
https://transformer-circuits.pub/2025/attribution-graphs/biology.html
Yannic Kilcher has a YouTube (part 1 of 2 parts is out so far):
https://www.youtube.com/watch?v=mU3g2YPKlsA
Sabine Hossenfelder has a YouTube:
https://www.youtube.com/watch?v=-wzOetb-D3w
Circuit Tracing: Revealing Computational Graphs in Language Models
https://transformer-circuits.pub/2025/attribution-graphs/methods.html
If you prefer reading a PDF version, try: https://webtopdf.com/
Additional background reading:
Faith and Fate: Limits of Transformers on Compositionality
https://arxiv.org/abs/2305.18654
On Limitations of the Transformer Architecture, Chapter 3 - The Impossibility of Composition
https://arxiv.org/abs/2402.08164
Fractal Generative Models
https://arxiv.org/pdf/2502.17437
YouTube:
https://www.youtube.com/watch?v=yxNuUg3aUjA
Github:
https://github.com/LTH14/fractalgen
From superposition to sparse codes: interpretable representations in neural networks
https://arxiv.org/pdf/2503.01824
There is at least one YouTube:
https://www.youtube.com/watch?v=t_i2NRr2eZA
u-µP: The Unit-Scaled Maximal Update Parametrization
https://arxiv.org/pdf/2407.17465
The Ultra-Scale Playbook: Training LLMs on GPU Clusters
https://huggingface.co/spaces/nanotron/ultrascale-playbook
There are a number of YouTubes on this.
MONA: Myopic Optimization with Non-myopic Approval Can Mitigate Multi-step Reward Hacking (Deepmind)
https://arxiv.org/pdf/2501.13011
There is a blog:
https://deepmindsafetyresearch.medium.com/mona-a-method-for-addressing-multi-step-reward-hacking-a31ac4b16483
There is at least one YouTube:
https://www.youtube.com/watch?v=mwqgIF3Ey8k
Scaling up Test-Time Compute with Latent Reasoning: A Recurrent Depth Approach
https://arxiv.org/pdf/2502.05171
s1: Simple test-time scaling
https://arxiv.org/abs/2501.19393
Github:
https://github.com/simplescaling/s1
There are many YouTubes, including:
https://www.youtube.com/watch?v=3tM3yc9UI84
and that YouTube mentions three similar papers published on almost the same date as the S1 paper:
Think Smarter not Harder: Adaptive Reasoning with Inference Aware Optimization (by Meta)
https://arxiv.org/abs/2501.17974
Large Language Models Think Too Fast To Explore Effectively (Georgia Institute of Tech.)
https://arxiv.org/pdf/2501.18009
Thoughts Are All Over the Place: On the Underthinking of o1-Like LLMs (TenCent AI Labs)
https://arxiv.org/pdf/2501.18585
Large Concept Models: Language Modeling in a Sentence Representation Space
https://ai.meta.com/research/publications/large-concept-models-language-modeling-in-a-sentence-representation-space/
A YouTube (many others):
https://www.youtube.com/watch?v=TwLiNTYvpPo
Deepseek R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
https://github.com/deepseek-ai/DeepSeek-R1/blob/main/DeepSeek_R1.pdf
There are many YouTubes and lots of press coverage
Base tech for R1 / background info / may also discuss if time:
Deepseek-V3 Technical Report
https://github.com/deepseek-ai/DeepSeek-V3/blob/main/DeepSeek_V3.pdf
Other R1-related info:
Berkeley Researchers Replicate Deepseek R1's Core Tech for Just $30
https://xyzlabs.substack.com/p/berkeley-researchers-replicate-deepseek
Jiayi Pan's discussion of what he and his team did:
https://x.com/jiayi_pirate/status/1882839370505621655
Berkeley team's code:
https://github.com/Jiayi-Pan/TinyZero
READ THIS FIRST:
This is a long paper (68 pages).
They are doing some cool, non-standard stuff with transformers.
That will be the focus of our discussion.
The "assigned reading" is Architecture, pages 21-37 (first part of Appendix), including Algorithms & Figures.
Skim the rest of the paper, as needed, to understand their context / what they are trying to do.
We may also look at their GitHub code, so you may want to take a look at that also.
---
Paper (focus on pages 21-37 - see the READ THIS above):
Simulating 500 million years of evolution with a language model
https://www.biorxiv.org/content/10.1101/2024.07.01.600583v2 <-- note v2 at end of URL
Github for model (open source):
https://github.com/evolutionaryscale/esm
YouTube by paper author:
https://www.youtube.com/watch?v=qeqbm8a1-ZA
Project page:
ESM3: Simulating 500 million years of evolution with a language model
https://www.evolutionaryscale.ai/blog/esm3-release
Huggingface for weights (open source license for non-commercial use; commercial use requires license):
https://huggingface.co/EvolutionaryScale/esm3
Nash Learning from Human Feedback
https://arxiv.org/pdf/2312.00886
TokenFormer: Rethinking Transformer Scaling with Tokenized Model Parameters
https://arxiv.org/pdf/2410.23168
There are many YouTubes including by Yannic Kilcher:
https://www.youtube.com/watch?v=gfU5y7qCxF0
and Gabriel Mongaras:
https://www.youtube.com/watch?v=4lGgbkD6Z0I
Visual Autoregressive Modeling: Scalable Image Generation via Next-Scale Prediction
https://arxiv.org/pdf/2404.02905
There is at least one YouTube:
https://www.youtube.com/watch?v=yJ396Ksiv2s
Generative Reward Models
https://arxiv.org/abs/2410.12832
Grokked Transformers are Implicit Reasoners: A Mechanistic Journey to the Edge of Generalization
https://arxiv.org/pdf/2405.15071
Github:
https://github.com/OSU-NLP-Group/GrokkedTransformer
OpenReview:
https://openreview.net/forum?id=ns8IH5Sn5y
Enhancing LLM Reasoning with Reward-guided Tree Search
https://arxiv.org/abs/2411.11694
Discrete Diffusion Modeling by Estimating the Ratios of the Data Distribution
https://arxiv.org/abs/2310.16834
YouTube (shorter):
https://www.youtube.com/watch?v=K_9wQ6LZNpI
YouTube (longer, by primary paper author):
https://www.youtube.com/watch?v=_1qv_LNjH9U
Github:
https://github.com/louaaron/Score-Entropy-Discrete-Diffusion
We will continue the discussion of:
The Llama 3 Herd of Models
https://arxiv.org/abs/2407.21783
We will start the discussion with a focus on Sections 7 and 8 (which we didn't have time for last week).
If time permits (it likely will) we will discuss (this week's new reading "assignment"):
- Section 3.3 through end of Section 3.3.4 (~pages 8 - 14)
- Section 6 (all of it) (~pages 51 - 53)
- Section 5 (skim for what whatever results catch your interest) (~pages 28 - 51)
- Any Figures and Tables that are referenced in the above readings.
- Anything anywhere in the paper that you want to discuss.
There are multiple YouTubes on the paper.
The Llama 3 Herd of Models
https://arxiv.org/abs/2407.21783
This is a long paper (92 pg) so we are skipping the sections on hardware, inference and results (leaves ~30 pg to read).
Our focus is on the software and architecture, including multi-modal aspects (the "assignment").
At the meetup we will discuss the paper, not read through it. Bring your questions, comments, etc.
Anyone is welcome to attend and listen without reading the "assignment".
If nobody reads it, the meeting will be short.
On the other hand, feel free to read more than the "assignment" and to share your wider insights in the meeting!
Here is the "assigned" reading with precise Sections shown:
- From the start through end of Section 3.2.1 (~pages 1 - 8)
- Section 3.4 through end of Section 4.3.7 (~pages 14-28)
- Section 7 through end of Section 7.5.7 (~pages 54-61)
- Section 8 through end of Section 8.3.2 (~pages 63-66)
- Any Figures and Tables that are referenced in the above readings.
A copy of the paper with the above sections marked is in this Github here:
https://github.com/davidmacmillan/DeepLearningStudyGroup/blob/master/The%20Llama%203%20Herd%20of%20Models%202407.21783v2.pdf
There are multiple YouTubes on the paper.
Agent S: An Open Agentic Framework that Uses Computers Like a Human
https://arxiv.org/pdf/2410.08164v1
Github:
https://github.com/simular-ai/Agent-S
There are a number of YouTubes on this paper.
nGPT: Normalized Transformer with Representation Learning on the Hypersphere
https://arxiv.org/pdf/2410.01131
There appear to be multiple YouTubes.
Open discussion of AI coding assist & AI coding completion tools people have used, and their assessment of them.
Interested in the full range of people's experiences with AI code tools: for code creation, code completion (copiloting), code debugging, and code refactoring.
Examples of code welcome but not required.
Diffusion Models are Evolutionary Algorithms
https://arxiv.org/pdf/2410.02543
Tweet:
https://x.com/YanboZhang3/status/1843134007892176995
Github:
https://github.com/Zhangyanbo/diffusion-evolution
At least one YouTube:
https://www.youtube.com/watch?v=Dh9gtg6N79U
Scaling Scaling Laws with Board Games
https://arxiv.org/pdf/2104.03113
Graph Retrieval-Augmented Generation: A Survey
https://arxiv.org/abs/2408.08921
YouTube (in Mandarin) (but click CC, then the Gear, then subtitles, then English):
https://www.youtube.com/watch?v=1OsVlbhMkek
Writing in the Margins: Better Inference Pattern for Long Context Retrieval
https://www.arxiv.org/abs/2408.14906
Diffusion Models Learn Low-Dimensional Distributions via Subspace Clustering∗
https://www.arxiv.org/abs/2409.02426
Paper for September 10, 2024
Unexpected Benefits of Self-Modeling in Neural Systems
https://arxiv.org/pdf/2407.10188
YouTube video
https://www.youtube.com/watch?v=yvHZ0nk8O5I
We are continuing the discussion of the paper from August 20, 2024:
The Remarkable Robustness of LLMs: Stages of Inference?
https://arxiv.org/abs/2406.19384
*** This is a long paper! Focus on pages 1-21 and skim Appendix D.8 as a representative output. ***
The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery
https://arxiv.org/pdf/2408.06292
Enticing or disturbing tweet:
https://x.com/Simeon_Cps/status/1823207094318735527
Blog:
https://sakana.ai/ai-scientist/
Github:
https://github.com/SakanaAI/AI-Scientist
The Remarkable Robustness of LLMs: Stages of Inference?
https://arxiv.org/abs/2406.19384
Segment 2 Anything (arXiv version):
https://arxiv.org/abs/2408.00714
Additional resources - Meta's Blog:
https://ai.meta.com/sam2/
Meta's Interactive Demo:
https://sam2.metademolab.com/
Meta's Announcement:
https://ai.meta.com/research/publications/sam-2-segment-anything-in-images-and-videos/
Github:
https://github.com/facebookresearch/segment-anything-2
There are a number of YouTube videos.
TextGrad: Automatic "Differentiation" via Text
https://arxiv.org/abs/2406.07496
Github:
https://github.com/zou-group/textgrad
Many YouTubes including:
https://youtu.be/Qks4UEsRwl0
The paper for July 30, 2024 is:
DETRs Beat YOLOs on Real-time Object Detection
https://arxiv.org/abs/2304.08069
Additional Background Materials - Project page:
https://zhao-yian.github.io/RTDETR/
Video (demo only):
https://www.youtube.com/watch?v=TbaLWroPYbo
Github:
https://github.com/lyuwenyu/RT-DETR
Background video on normal DETR by Meta (creator also has videos on other object detection models):
https://www.youtube.com/watch?v=A2f4w54fSsM
When Can LLMs Actually Correct Their Own Mistakes? A Critical Survey of Self-Correction of LLMs
https://arxiv.org/abs/2406.01297
xLSTM: Extended Long Short-Term Memory
https://arxiv.org/abs/2405.04517
YouTube (Yannic Kilcher)
https://www.youtube.com/watch?v=0OaEv1a5jUM
YouTube (Gabriel Mongaras)
https://www.youtube.com/watch?v=4ND8lU2aN_k
Medium article
https://medium.com/@AIBites/xlstm-extended-long-short-term-memory-networks-c4ba34fdd98d
6:30 PM - face-to-face get together & casual sit-down dinner. No paper this week.
At: Agave Mexican Bistro, 194 Castro Street, Mountain View, California 94041.
Paper: Banishing LLM Hallucinations Requires Rethinking Generalization
https://arxiv.org/abs/2406.17642
Github:
https://github.com/lamini-ai/
Transformers are RNNs: Fast Autoregressive Transformers with Linear Attention
https://arxiv.org/abs/2006.16236
YouTube by Yannic Kilcher on paper (may be others):
https://www.youtube.com/watch?v=hAooAOFRsYc
Transformers are SSMs: Generalized Models and Efficient Algorithms Through Structured State Space Duality (Mamba 2)
https://arxiv.org/abs/2405.21060
This blog:
https://gonzoml.substack.com/p/mamba-2-is-here
and the 4 referenced blogs starting here:
https://goombalab.github.io/blog/2024/mamba2-part1-model/
are more approachable.
"Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet"
https://transformer-circuits.pub/2024/scaling-monosemanticity/index.html
Tweet thread / overview & highlights:
https://x.com/mlpowered/status/1792948212728524917
"Scaling Monosemanticity: Extracting Interpretable Features from Claude 3 Sonnet"
https://transformer-circuits.pub/2024/scaling-monosemanticity/index.html
Tweet thread / overview & highlights:
https://x.com/mlpowered/status/1792948212728524917
Good video on this week's paper (blog):
https://www.youtube.com/watch?v=y0ZXFl3rQlQ
The Platonic Representation Hypothesis
https://arxiv.org/pdf/2405.07987
Github / project page:
https://phillipi.github.io/prh/
Github with their code:
https://github.com/minyoungg/platonic-rep
There are also a number of YouTubes that discuss the paper.
Instead of a paper, we are going to go through Andrej Karpathy's YouTube video on creating transformer code:
https://www.youtube.com/watch?v=kCc8FmEb1nY
Have the colab or github code loaded on your PC before, ready to go through, so you don't have to type it in during our session.
Colab:
https://colab.research.google.com/drive/1JMLa53HDuA-i7ZBmqV7ZnA3c_fvtXnx-
Github:
https://github.com/karpathy/ng-video-lecture
KAN: Kolmogorov-Arnold Networks
https://arxiv.org/pdf/2404.19756
iTransformer: Inverted Transformers Are Effective for Time Series Forecasting
https://arxiv.org/pdf/2310.06625.pdf
Chronos: Learning the Language of Time Series
https://arxiv.org/abs/2403.07815
There is a YouTube on the paper:
https://www.youtube.com/watch?v=yKKWCqABspw
Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention
https://arxiv.org/pdf/2404.07143.pdf
From DeepMind, on their generalized AI that can play arbitrary video games,
Scalable Instructable Multiworld Agent (SIMA AI).
Paper: Scaling Instructable Agents Across Many Simulated Worlds
https://storage.googleapis.com/deepmind-media/DeepMind.com/Blog/sima-generalist-ai-agent-for-3d-virtual-environments/Scaling%20Instructable%20Agents%20Across%20Many%20Simulated%20Worlds.pdf
Additional resources:
Deepmind blog:
https://deepmind.google/discover/blog/sima-generalist-ai-agent-for-3d-virtual-environments/
2 minute paper:
https://www.youtube.com/watch?v=5U_Q2Lmnq_c
Longer YouTube:
https://www.youtube.com/watch?v=ymKkfRu6dz4
Cancelled
Evolutionary Optimization of Model Merging Recipes
https://arxiv.org/pdf/2403.13187.pdf
Solving Olympiad Geometry Without Human Demonstrations
https://www.nature.com/articles/s41586-023-06747-5
There are a number of YouTubes, including:
https://www.youtube.com/watch?v=ZobxevIJQ7A
and (Yannic)
https://www.youtube.com/watch?v=ZNK4nfgNQpM
Interpretability in the Wild: a Circuit for Indirect Object Identification in GPT-2 small
https://arxiv.org/abs/2211.00593
Additional material - YouTube with the authors (in 2 parts):
https://www.youtube.com/watch?v=gzwj0jWbvbo
and
https://www.youtube.com/watch?v=b9xfYBKIaX4
Still want more? Two more YouTubes by the YouTube channel owner, Neel Nanda, on his research, inspired by this week's paper.
(Neel is head of DeepMind's interoperability team - see https://www.neelnanda.io/about)
https://www.youtube.com/watch?v=m8tzXelUTLo
and
https://www.youtube.com/watch?v=tiHRceW-19U
A Review of Sparse Expert Models in Deep Learning
https://arxiv.org/pdf/2209.01667.pdf
Background paper:
Twenty Years of Mixture of Experts
https://www.ee.hacettepe.edu.tr/~eyuksel/Publications/2012_TwentyYearsofMixtureofExperts.pdf
Look Before You Leap: A Universal Emergent Decomposition of Retrieval Tasks in Language Models
https://arxiv.org/abs/2312.10091
Representation Engineering draws on insights from cognitive neuroscience to engineer neural representations, rather than neurons or circuits. Rep. Eng. can be used to apply a control vector during inference to change or limit a model's behavior.
Paper:
Representation Engineering - a Top-Down Approach to AI Transparency
https://arxiv.org/pdf/2310.01405.pdf
Additional background info:
Blog:
Representation Engineering Mistral-7B an Acid Trip
https://vgel.me/posts/representation-engineering/
Another blog:
Steering GPT-2-XL by adding an activation vector
https://www.lesswrong.com/posts/5spBue2z2tw4JuDCx/steering-gpt-2-xl-by-adding-an-activation-vector
Third blog:
https://www.astralcodexten.com/p/the-road-to-honest-ai
Github:
https://github.com/andyzoujm/representation-engineering
Github - Python library
https://github.com/vgel/repeng/
Grandmaster-Level Chess Without Search
https://arxiv.org/abs/2402.04494
Mistral 7B
https://arxiv.org/pdf/2310.06825.pdf
Mixtral of Experts
https://arxiv.org/pdf/2401.04088.pdf
Optional:
There are many YouTubes on each, including by Yannic Kilcher.
There are download-and-run llamafile quantized versions of Mistral 7B and Mixtral 8x7B at:
https://github.com/Mozilla-Ocho/llamafile
(Mac and Linux, Windows has a few very minor additional steps.)
Why think step by step - Reasoning emerges from the locality of experience
https://arxiv.org/pdf/2304.03843.pdf
Direct Preference Optimization: Your Language Model is Secretly a Reward Model
https://arxiv.org/pdf/2305.18290.pdf
Optional:
There are lots of YouTubes on DPO to choose from.
A related github by lucidrains:
https://github.com/lucidrains/self-rewarding-lm-pytorch
A couple of more recent related papers:
Self-Play Fine-Tuning Converts Weak Language Models to Strong Language Models
https://arxiv.org/pdf/2401.01335v1.pdf
and
Self-Rewarding Language Models
https://arxiv.org/pdf/2401.10020.pdf
"Latent Consistency Models: Synthesizing High-Resolution Images with Few-Step Inference
https://arxiv.org/pdf/2310.04378.pdf
Additional background items:
There is at least one YouTube on this paper:
https://www.youtube.com/watch?v=OT3JWNz0Il8
Huggingface demos:
https://huggingface.co/collections/latent-consistency/latent-consistency-model-demos-654e90c52adb0688a0acbe6f
LCM-LoRA: A Universal Stable-Diffusion Acceleration Module
https://arxiv.org/abs/2311.05556
Consistency Models https://arxiv.org/abs/2303.01469
There are also multiple YouTubes on Consistency Models.
Mamba: Linear-Time Sequence Modeling with Selective State Spaces
https://arxiv.org/ftp/arxiv/papers/2312/2312.00752.pdf
A few optional videos (likely are others too):
Video: https://youtu.be/ouF-H35atOY?si=BFQ_PTVfhfNXBLPb
Video: https://www.youtube.com/watch?v=ouF-H35atOY
The Clock and the Pizza: Two Stories in Mechanistic Explanation of Neural Networks
https://arxiv.org/pdf/2306.17844.pdf
paper: MemGPT -Towards LLMs as an Operating System https://arxiv.org/pdf/2310.08560.pdf
Blog w MemBPT - https://memgpt.ai/
youtube: https://www.youtube.com/watch?v=nQmZmFERmrg
paper:
CLUSTERFORMER: Clustering As A Universal Visual Learner
https://openreview.net/pdf?id=S1KGaTSOTS
paper:
An Emulator for Fine-Tuning Large Language Models using Small Language Models
https://arxiv.org/pdf/2310.12962.pdf
paper:
From attribution maps to human-understandable explanations through Concept Relevance Propagation
https://www.nature.com/articles/s42256-023-00711-8
paper:
Liquid Structural State-Space Models
https://arxiv.org/pdf/2209.12951.pdf
paper:
Liquid Time-Constant Networks
https://arxiv.org/abs/2006.04439
youtube:
https://www.youtube.com/watch?v=IlliqYiRhMU
shorter video:
https://www.youtube.com/watch?v=RI35E5ewBuI
paper
3D Gaussian Splatting for Real-Time Radiance Field Rendering
https://arxiv.org/abs/2308.04079
youtubes:
Superb 2 minute video on paper
https://www.youtube.com/watch?v=HVv_IQKlafQ
Siggraph 2023 talk on paper - this is 5 minutes
https://www.youtube.com/watch?v=T_kXY43VZnk&t=3s
Author's blog, including links to code:
https://repo-sam.inria.fr/fungraph/3d-gaussian-splatting/
paper: https://arxiv.org/abs/2112.04035
Relating transformers to models and neural representations of the hippocampal formation
another paper:
https://amygdala.psychdept.arizona.edu/labspace/JclubLabMeetings/JeanMarc-Build-cognitive-maps.pdf -
How to build a cognitive map
YouTubes:
How Your Brain Organizes Information
https://www.youtube.com/watch?v=9qOaII_PzGY&t=413s
Can We Build an Artificial Hippocampus?
https://www.youtube.com/watch?v=cufOEzoVMVA
The Tolman-Eichenbaum Machine: Unifying Space and Relational Memory through Generalization in the Hippocampal Formation
https://www.cell.com/cell/fulltext/S0092-8674(20)31388-X
paper:
3D Gaussian Splatting for Real-Time Radiance Field Rendering
https://research.nvidia.com/labs/par/Perfusion/
paper:
Imagic: Text-Based Real Image Editing with Diffusion Models
https://arxiv.org/pdf/2210.09276.pdf
YouTube:
https://www.youtube.com/watch?v=PzHMjCtuPuo
blog:
https://imagic-editing.github.io/
LongNet: Scaling Transformers to 1,000,000,000 Tokens
paper: https://arxiv.org/abs/2307.02486
Blog:
https://syncedreview.com/2023/07/10/microsofts-longnet-scales-transformer-to-one-billion-tokens
Consciousness in Artificial Intelligence: Insights from the Science of Consciousness
https://arxiv.org/pdf/2308.08708.pdf
paper:
A Theory for Emergence of Complex Skills in Language Models and video
https://arxiv.org/pdf/2307.15936.pdf
youtube:
https://www.youtube.com/watch?v=0D23NeBjCeQ
Paper: Neural Laplace: Learning diverse classes of differential equations in the Laplace domain
https://arxiv.org/pdf/2206.04843.pdf
Slides and video from ICML 2022:
https://icml.cc/virtual/2022/oral/16728
paper: https://arxiv.org/abs/2308.03296 - Studying Large Language Model Generalization with Influence Functions
blog: https://www.anthropic.com/index/influence-functions
paper: Music Generations https://arxiv.org/pdf/2306.05284.pdf
blog: https://about.fb.com/news/2023/08/audiocraft-generative-ai-for-music-and-audio/
blog: https://ai.meta.com/blog/audiocraft-musicgen-audiogen-encodec-generative-ai-audio/
paper: https://arxiv.org/abs/2205.10343 Towards Understanding Grokking: An Effective Theory of Representation Learning
blog: https://ericjmichaud.com/grokking-squared/
blog: https://www.beren.io/2022-01-11-Grokking-Grokking/
blog: https://www.beren.io/2022-04-17-Understanding_Overparametrized_Generalization/
paper: Mixture of experts (similar to chatGPT4): https://arxiv.org/abs/2305.14705
blog: Mixture-of-Experts with Expert Choice Routing -
https://ai.googleblog.com/2022/11/mixture-of-experts-with-expert-choice.html
blot: Introducing Pathways: A next-generation AI architecture
https://blog.google/technology/ai/introducing-pathways-next-generation-ai-architecture/
We're going to cover Chapter 16 Deep Networks for Classification from the following book:
https://book-wright-ma.github.io/Book-WM-20210422.pdf - High dimensional Data Analysis with Low Dimensional Models
blog: https://terrytao.wordpress.com/2007/04/13/compressed-sensing-and-single-pixel-cameras/#more-25
We're going to cover the 4th chapter of this book.
https://book-wright-ma.github.io/Book-WM-20210422.pdf - High dimensional Data Analysis with Low Dimensional Models
We're going to cover the 1st chapter of this book.
https://book-wright-ma.github.io/Book-WM-20210422.pdf - High dimensional Data Analysis with Low Dimensional Models
Blog: https://terrytao.wordpress.com/2007/04/13/compressed-sensing-and-single-pixel-cameras/#more-25
paper: https://arxiv.org/pdf/2305.17126.pdf - Large Language Models as Tool Makers
youtube: https://www.youtube.com/watch?v=qWI1AJ2nSDY
youtube: https://www.youtube.com/watch?v=KXlPzMRTfMk
youtube: https://www.youtube.com/watch?v=srDVNbxPgZI
Consciousness as a Memory System https://pubmed.ncbi.nlm.nih.gov/36178498/
https://arxiv.org/abs/1804.08838
Blog: https://www.uber.com/blog/intrinsic-dimension/
more good stuff on intrinsic dimension:
Nature paper: https://www.nature.com/articles/s41598-017-11873-y
Wikipedia: https://en.wikipedia.org/wiki/Intrinsic_dimension
Application - Yann LeCun at 57:15 on does text fully represent world model?
https://www.youtube.com/watch?v=SGzMElJ11Cc
vs. differing view from Ilya Sutskever at 15:30
https://www.youtube.com/watch?v=SjhIlw3Iffs
Applying intrinsic dimension to scaling laws in training / loss:
https://jmlr.csail.mit.edu/papers/volume23/20-1111/20-1111.pdf
https://arxiv.org/abs/2102.06701
Paper: https://arxiv.org/pdf/2305.16291.pdf
Twit: Tweet with nice overview by author https://twitter.com/DrJimFan/status/1662117784023883777
Code: https://github.com/MineDojo/Voyager
website: https://voyager.minedojo.org/
paper: https://arxiv.org/pdf/2203.15556.pdf - Training Compute-Optimal Large Language Models
blog: https://www.lesswrong.com/posts/6Fpvch8RR29qLEWNH/chinchilla-s-wild-implications
blog: https://www.harmdevries.com/post/model-size-vs-compute-overhead/
google blog: https://www.cnbc.com/2023/05/16/googles-palm-2-uses-nearly-five-times-more-text-data-than-predecessor.html
paper: https://arxiv.org/abs/2212.09720 - The case for 4-bit precision: k-bit Inference Scaling Laws
paper: https://arxiv.org/pdf/2210.17323.pdf - GPTQ: ACCURATE POST-TRAINING QUANTIZATION FOR GENERATIVE PRE-TRAINED TRANSFORMERS
paper: https://arxiv.org/pdf/2106.09685.pdf - LORA: LOW-RANK ADAPTATION OF LARGE LANGUAGE MODELS
paper: https://arxiv.org/pdf/2210.03629.pdf - REACT: SYNERGIZING REASONING AND ACTING IN LANGUAGE MODELS
paper: https://www.pinecone.io/learn/locality-sensitive-hashing/
paper: https://arxiv.org/pdf/2201.11903.pdf - Chain of thought prompting elicits reasoning in large language models.
paper: https://arxiv.org/pdf/2210.03629.pdf - REACT: SYNERGIZING REASONING AND ACTING IN LANGUAGE MODELS
paper: https://www.pinecone.io/learn/locality-sensitive-hashing/
https://python.langchain.com/en/latest/modules/agents.html
https://arxiv.org/pdf/2210.03629.pdf - REACT: SYNERGIZING REASONING AND ACTING IN LANGUAGE MODELS
https://www.pinecone.io/learn/locality-sensitive-hashing/
Blog: https://yoheinakajima.com/task-driven-autonomous-agent-utilizing-gpt-4-pinecone-and-langchain-for-diverse-applications/
Code: https://github.com/hwchase17/langchain
Paper: Eliciting Latent Predictions from Transformers with the Tuned Lens https://arxiv.org/abs/2303.08112
Paper: https://openreview.net/pdf?id=lMMaNf6oxKM - Recipe for a General, Powerful, Scalable Graph Transformer
youtube: https://www.youtube.com/watch?v=DiLSCReBaTg
Paper: https://proceedings.neurips.cc/paper/2021/hash/f1c1592588411002af340cbaedd6fc33-Abstract.html - Do Transformers Really Perform Badly for Graph Representation?
video: https://www.youtube.com/watch?v=FKuQpPIRjLk - review by authors
video: https://www.youtube.com/watch?v=xQ5ltOOxoFg
Paper: https://arxiv.org/abs/2212.07359 - Post-hoc Uncertainty Learning using a Dirichlet Meta-Model
youtube: https://www.youtube.com/watch?v=nE8XJ1f0zO0
Paper: https://arxiv.org/abs/2202.05262 - Locating and Editing Factual Associations in GPT
blog: https://rome.baulab.info/
Yannic video: https://www.youtube.com/watch?v=_NMQyOu2HTo
Paper: Human-Timescale Adaptation in an Open-Ended Task Space: https://arxiv.org/pdf/2301.07608.pdf
https://www.youtube.com/watch?v=A2hOWShiYoM
https://sites.google.com/view/adaptive-agent/
Paper: Toolformer: Language Models Can Teach Themselves to Use Tools: https://arxiv.org/abs/2302.04761
Paper: https://arxiv.org/pdf/2203.02155.pdf - Training language models to follow instructions with human feedback
Paper: https://arxiv.org/pdf/2111.15664.pdf - OCR-free Document Understanding Transformer
Paper: https://arxiv.org/abs/2205.06175 - A generalist agent - Gato
YouTube: Eden Mayer https://www.youtube.com/watch?v=wSQJZHfAg18
YouTube - Jay Alamar https://www.youtube.com/watch?v=kT6DYKgWNHg
YouTube - Lex Fridman and Oriol Vinyals on How Gato Works https://www.youtube.com/watch?v=vwB9zO2h9j0
Overview - main site on Gato at Deepmind https://www.deepmind.com/publications/a-generalist-agent
blog review - https://arshren.medium.com/deep-minds-generalist-agent-gato-209969e12782
Paper: https://openreview.net/pdf?id=M95oDwJXayG - ADDRESSING PARAMETER CHOICE ISSUES IN UNSUPERVISED DOMAIN ADAPTATION BY AGGREGATION
Paper: https://arxiv.org/pdf/2301.04104v1.pdf - Mastering Diverse Domains through World Models
Blog: https://danijar.com/project/dreamerv3/
YouTube: https://www.youtube.com/watch?v=vfpZu0R1s1Y
Paper: https://arxiv.org/abs/2212.04089 - Composable NN: Editing Models With Task Arithmetic
Paper: https://arxiv.org/pdf/1707.06690.pdf - DeepPath: A Reinforcement Learning Method for Knowledge Graph Reasoning
Paper: https://arxiv.org/abs/2212.04458 - GENERAL-PURPOSE IN-CONTEXT LEARNING BY META-LEARNING TRANSFORMERS
paper: https://arxiv.org/pdf/2209.04836.pdf - Git Re-Basin: Merging Models modulo Permutation Symmetries
paper: https://arxiv.org/abs/2012.09855 - Infinite Nature: Perpetual View Generation of Natural Scenes from a Single Image
blog: https://infinite-nature.github.io/
Paper: https://arxiv.org/abs/2206.00364 - Elucidating the Design Space of Diffusion-Based Generative Models
video: https://www.youtube.com/watch?v=OYiQctx7kDE
paper: https://arxiv.org/pdf/2206.10991.pdf - Graph Neural Networks as Gradient Flows: understanding graph convolutions via energy
youtube (author): https://www.youtube.com/watch?v=sgTTtmwOMgE
youtube: https://www.youtube.com/watch?v=hmI4C6AodEQ
paper: https://www.pnas.org/doi/full/10.1073/pnas.2016239118
video: https://slideslive.com/38942412/biological-structure-and-function-emerge-from-scaling-unsupervised-learning-to-250-million-protein-sequences
paper: https://arxiv.org/pdf/2209.11178.pdf - Poisson Flow Generative Models
paper: https://arxiv.org/pdf/2209.12892.pdf - LEARNING TO LEARN WITH GENERATIVE MODELS OF NEURAL NETWORK CHECKPOINTS
blog: https://www.marktechpost.com/2022/10/21/latest-machine-learning-research-at-uc-berkeley-proposes-a-way-to-design-a-learned-optimizer-using-generative-models-of-neural-network-checkpoints/
author blog: https://www.wpeebles.com/Gpt.html
paper: Cellular automata as convolutional neural networks https://arxiv.org/pdf/1809.02942.pdf
survey: Collective Intelligence for Deep Learning: A Survey of Recent Developments https://arxiv.org/abs/2111.14377
demo: Self-classifying MNIST Digits https://distill.pub/2020/selforg/mnist/
paper: https://proceedings.mlr.press/v162/zhu22c/zhu22c.pdf - Neural-Symbolic Models for Logical Queries on Knowledge Graphs
paper: https://arxiv.org/pdf/2206.02768.pdf - The Neural Covariance SDE: Shaped Infinite Depth-and-Width Networks at Initialization
paper: https://papers.nips.cc/paper/2019/file/952285b9b7e7a1be5aa7849f32ffff05-Paper.pdf - Legendre Memory Units: Continuous-Time
paper: https://arxiv.org/pdf/2208.01618.pdf - An Image is Worth One Word: Personalizing Text-to-Image Generation using Textual Inversion
githup.io: https://textual-inversion.github.io/
YouTube https://www.youtube.com/watch?v=f3oXa7_SYek
paper: https://arxiv.org/pdf/2205.14415.pdf - Non-stationary Transformers: Rethinking the Stationarity in Time Series Forecasting
paper: https://arxiv.org/abs/2110.02402 - Language Modeling using LMUs: 10x Better Data Efficiency or Improved Scaling Compared to Transformers
youtube vid: https://www.youtube.com/watch?v=8t64QaTdBcU
Paper: HOW NEURAL NETWORKS EXTRAPOLATE: FROM FEEDFORWARD TO GRAPH NEURAL NETWORKS - https://arxiv.org/pdf/2009.11848.pdf
Paper: Masked Siamese Networks for Label-Efficient Learning - https://arxiv.org/abs/2204.07141
Paper: Principle of Maximal Coding Rate Reduction https://arxiv.org/abs/2006.08558
ReduNet: https://arxiv.org/pdf/2105.10446.pdf
Github: https://github.com/ryanchankh/mcr2
Paper: On the Principles of Parsimony and Self-Consistency for the Emergence of Intelligence https://arxiv.org/abs/2207.04630
Background: On the Principles of Parsimony and Self-Consistency for the Emergence of Intelligence https://arxiv.org/abs/2207.04630
Background: https://www.youtube.com/watch?v=OIVcfZeR1CE youtube by author
Background: https://cmsa.fas.harvard.edu/wp-content/uploads/2021/04/Deep_Networks_from_First_Principles.pdf - slides by author
Paper: Data Distributional Properties Drive Emergent In-Context Learning in Transformers https://arxiv.org/pdf/2205.05055.pdf
Paper: A Mathematical Framework for Transformer Circuits https://transformer-circuits.pub/2021/framework/index.html#model-simplifications
Paper: A Mathematical Framework for Transformer Circuits https://transformer-circuits.pub/2021/framework/index.html#model-simplifications
Paper: https://arxiv.org/abs/2001.08361 - Scaling Laws for Neural Language Models
Blog: https://medium.com/nlplanet/two-minutes-nlp-scaling-laws-for-neural-language-models-add6061aece7
Paper: https://arxiv.org/abs/2206.11795 - Video PreTraining (VPT): Learning to Act by Watching Unlabeled Online Videos
https://github.com/openai/Video-Pre-Training
Yannic Review: https://www.youtube.com/watch?v=oz5yZc9ULAc
Paper: https://arxiv.org/pdf/2110.00966.pdf - Translating Images into Maps
Paper: https://arxiv.org/abs/2205.09665 - Automated Crossword Solving
Paper: https://arxiv.org/pdf/2205.10824.pdf - ReLU Fields: The Little Non-linearity That Could
Paper: https://arxiv.org/abs/2102.06810 - Understanding Self-Supervised Learning Dynamics without Contrastive Pairs
Paper: https://arxiv.org/pdf/2205.06175.pdf - A Generalist Agent
Blog: https://www.deepmind.com/publications/a-generalist-agent
https://arxiv.org/pdf/2202.05780.pdf - A Modern Self-Referential Weight Matrix That Learns to Modify Itself
https://openreview.net/pdf?id=M752z9FKJP - LEARNING STRIDES IN CONVOLUTIONAL NEURAL NETWORKS
https://openreview.net/pdf?id=b-ny3x071E5 - BOOTSTRAPPED META-LEARNING
https://arxiv.org/abs/2202.06991 - Transformer Memory as a Differentiable Search Index
https://www.youtube.com/watch?v=C7mUYocWdG0 - Yannic author interview
https://www.youtube.com/watch?v=qlB0TPBQ7YY - Yannic on Transformer paper
https://arxiv.org/abs/2204.06125 - Hierarchical Text-Conditional Image Generation with CLIP Latents
https://openai.com/dall-e-2/ - OpenAI blog
https://www.youtube.com/watch?v=j4xgkjWlfL4 - yannic video
https://arxiv.org/pdf/2103.00020.pdf - Learning Transferable Visual Models From Natural Language Supervision
https://www.youtube.com/watch?v=1LUWWAnK_Ks
https://www.youtube.com/watch?v=3X3EY2Fgp3g
https://arxiv.org/pdf/2110.13985.pdf - Combining Recurrent, Convolutional, and Continuous-time Models with Linear State-Space Layers
https://arxiv.org/pdf/2202.00666.pdf - Typical Decoding for Natural Language Generation
https://www.youtube.com/watch?v=AvHLJqtmQkE
https://arxiv.org/pdf/2105.04906.pdf - VICREG: VARIANCE-INVARIANCE-COVARIANCE REGULARIZATION FOR SELF-SUPERVISED LEARNING
https://www.youtube.com/watch?v=MzKDNmOJ67Q
https://openreview.net/forum?id=4orlVaC95Bo - Task-Agnostic Undesirable Feature Deactivation Using Out-of-Distribution Data
https://arxiv.org/abs/2203.03466 - Tensor Programs V: Tuning Large Neural Networks via Zero-Shot Hyperparameter Transfer
https://www.youtube.com/watch?v=MNOJQINH-qw
https://arxiv.org/abs/2201.12122 - Can Wikipedia Help Offline Reinforcement Learning?
Yannic's talk on this,
https://www.youtube.com/watch?v=XHGh19Hbx48
and he also has a followon video interview with the authors
https://www.youtube.com/watch?v=FNDVy_BR8aA
https://arxiv.org/pdf/2107.03342.pdf - A Survey of Uncertainty in Deep Neural Networks
https://arxiv.org/pdf/2201.08239v2.pdf - LaMDA: Language Models for Dialog Applications
https://openreview.net/pdf?id=TrjbxzRcnf- MEMORIZING TRANSFORMERS
https://arxiv.org/pdf/2106.07644.pdf - A Continuized View on Nesterov Acceleration for Stochastic Gradient Descent and Randomized Gossip
https://arxiv.org/pdf/2108.08052.pdf - Moser Flow: Divergence-based Generative Modeling on Manifolds
https://dylandoblar.github.io/noether-networks/ - Noether Networks: meta-learning useful conserved quantities
https://www.youtube.com/watch?v=Xp3jR-ttMfo
https://arxiv.org/pdf/2010.15277.pdf - Class-incremental learning: survey and performance evaluation on image classification
https://arxiv.org/abs/2006.11287 - Discovering Symbolic Models from Deep Learning with Inductive Biases
https://arxiv.org/pdf/2006.09252.pdf - Improving Graph Neural Network Expressivity via Subgraph Isomorphism Counting
https://arxiv.org/pdf/2112.04426.pdf - Improving Language Models by Retrieving from Trillions of Tokens
https://arxiv.org/abs/2106.01798 - Implicit MLE: Backpropagating Through Discrete Exponential Family Distributions
https://www.youtube.com/watch?v=W2UT8NjUqrk
https://arxiv.org/pdf/2108.01073.pdf - Image Synthesis and Editing with Stochastic Differential Equations
https://openreview.net/forum?id=HfpNVDg3ExA OpenReviewOpenReview Probabilistic Transformer For Time Series Analysis
https://arxiv.org/pdf/2110.03922.pdf - NEURAL TANGENT KERNEL EIGENVALUES ACCURATELY PREDICT GENERALIZATION
https://arxiv.org/pdf/2104.00681.pdf - NeuralRecon: Real-Time Coherent 3D Reconstruction from Monocular Video
https://github.com/zju3dv/NeuralRecon
https://arxiv.org/pdf/2110.09485.pdf - Learning in High Dimension Always Amounts to Extrapolation
https://arxiv.org/pdf/2109.02355.pdf - A Farewell to the Bias-Variance Tradeoff? An Overview of the Theory of Overparameterized Machine Learning
https://arxiv.org/pdf/2006.09011.pdf - Improved Techniques for Training Score-Based Generative Models
https://arxiv.org/abs/2006.05929 - Dataset Condensation with Gradient Matching
https://arxiv.org/abs/1811.10959 - Dataset distillation
https://arxiv.org/pdf/2003.13216.pdf - Learning to Learn Single Domain Generalization
https://arxiv.org/pdf/2108.11482.pdf - ETA Prediction with Graph Neural Networks in Google Maps
https://cascaded-diffusion.github.io/assets/cascaded_diffusion.pdf - Cascaded Diffusion Models for High Fidelity Image Generation
https://arxiv.org/pdf/2107.06277.pdf - Why Generalization in RL is Difficult: Epistemic POMDPs and Implicit Partial Observability
https://arxiv.org/abs/2108.07732 - Program Synthesis with Large Models
https://arxiv.org/abs/2012.13349 - Solving Mixed Integer Programs Using Neural Networks
https://www.nature.com/articles/s41586-021-03819-2 - DeepFold
Alphafold - blog https://deepmind.com/blog/article/alphafold-a-solution-to-a-50-year-old-grand-challenge-in-biology paper https://www.nature.com/articles/s41586-021-03819-2 supplemental info https://static-content.springer.com/esm/art%3A10.1038%2Fs41586-021-03819-2/MediaObjects/41586_2021_3819_MOESM1_ESM.pdf
https://www.zdnet.com/article/googles-supermodel-deepmind-perceiver-is-a-step-on-the-road-to-an-ai-machine-that-could-process-everything/ https://arxiv.org/abs/2103.03206
https://arxiv.org/pdf/1503.03585.pdf (Deep Unsupervised Learning using Non equilibrium Thermodynamics) by Surya Ganguli at Stanford
Wednesday, July 7, 2021 - https://arxiv.org/pdf/2105.05233.pdf - Diffusion Models Beat GANs on Image Synthesis
https://arxiv.org/pdf/2006.11239.pdf - Denoising Diffusion Probabilistic Models
https://arxiv.org/abs/2010.03409 - Learning mesh-based simulation with graph networks
https://sites.google.com/view/learning-to-simulate
https://deepmind.com/research/publications/Learning-to-Simulate-Complex-Physics-with-Graph-Networks
https://arxiv.org/pdf/2106.01345.pdf - Decision Transformer: Reinforcement Learning via Sequence Modeling
https://www.youtube.com/watch?v=-buULmf7dec
https://sites.google.com/berkeley.edu/decision-transformer
https://arxiv.org/pdf/2103.07945.pdf - Learning One Representation to Optimize All Rewards
https://distill.pub/2021/multimodal-neurons/ - Multimodal Neurons in Artificial Neural Networks
https://openai.com/blog/clip/ - CLIP: Connecting Text and Images
https://arxiv.org/pdf/2104.14294.pdf - Emerging Properties in Self-Supervised Vision Transformers
https://arxiv.org/pdf/2104.10558.pdf - Contingencies from Observations: Tractable ContingencyPlanning with Learned Behavior Models
https://arxiv.org/pdf/1806.09055.pdf - DARTS: Differentiable Architecture Search (ICLR 2019)
https://arxiv.org/pdf/2104.06644.pdf - Masked Language Modeling and the Distributional Hypothesis:Order Word Matters Pre-training for Little
https://arxiv.org/pdf/2009.03717.pdf - Hierarchical message passing graph neural networks
https://arxiv.org/pdf/2103.03230v1.pdf - Barlow Twins: Self-Supervised Learning via Redundancy Reduction
https://arxiv.org/pdf/2103.14770.pdf - Categorical representation learning: morphism is all you need
https://arxiv.org/pdf/2102.12736v1.pdf - Time-Series Imputation with Wasserstein Interpolation for Optimal Look-Ahead-Bias and Variance Tradeoff
https://awacrl.github.io/ - Accelerating online reinforcement learning with offline datasets
https://arxiv.org/pdf/2102.12092.pdf - Zero-Shot Text-to-Image Generation
https://openai.com/blog/dall-e/
https://giotto-ai.github.io/gtda-docs/latest/notebooks/gravitational_waves_detection.html
https://arxiv.org/pdf/2102.08602.pdf - Modeling long-range interactions without attention
https://arxiv.org/pdf/2101.08692.pdf - Characterizing signal propagation to close the performance gap in unnormalized resnets
https://arxiv.org/pdf/2006.10742.pdf - Learning Invariant Representations forReinforcement Learning without Reconstruction
https://arxiv.org/pdf/2007.13544.pdf - Combining Deep Reinforcement Learning and Search for Imperfect-Information Games
https://arxiv.org/pdf/2010.11929.pdf - An image is worth 16x16 words: transformers for image recognition at scale
https://arxiv.org/abs/2003.02821 - What went wrong and when? Instance-wise feature importance for time-series black-box models
https://arxiv.org/pdf/1912.09363.pdf - Temporal Fusion Transformersfor Interpretable Multi-horizon Time Series Forecasting
https://arxiv.org/abs/1905.10403 - Neural Jump Stochastic Differential Equations
http://implicit-layers-tutorial.org/neural_odes/ - We're continuing this from last week. This week we'll cover Ch 3,4,5.
http://implicit-layers-tutorial.org/ - NeurIPS tutorial on deep implicit networks
https://arxiv.org/pdf/1907.03907.pdf - Latent ODEs for Irregularly-Sampled Time Series
https://www.youtube.com/watch?v=tOkH339Wucs
https://papers.nips.cc/paper/2020/file/08425b881bcde94a383cd258cea331be-Paper.pdf - Ridge Rider: Finding Diverse Solutions by FollowingEigenvectors of the Hessian
https://proceedings.neurips.cc/paper/2020/file/28e209b61a52482a0ae1cb9f5959c792-Paper.pdf “OOD-MAML: Meta-Learning for Few-Shot Out-of-Distribution Detection and Classification"
https://arxiv.org/pdf/2011.02421.pdf - ONE-SHOT CONDITIONAL AUDIO FILTERING OF ARBITRARY SOUNDS
https://arxiv.org/pdf/2010.14498.pdf - Implicit under-parametrization inhibits data efficient deep reinforcement learning
https://arxiv.org/pdf/2010.03759.pdf - Energy-based Out-of-distribution Detection
https://arxiv.org/abs/2005.01643 - offline reinforcement learning - tutorial review and perspectives on open problems
https://arxiv.org/pdf/2009.12981.pdf - Parametric UMAP: Learning embeddings with deep neural networks for representation and semi-supervised learning
https://arxiv.org/pdf/2009.12981.pdf - Parametric UMAP: Learning embeddings with deep neural networks for representation and semi-supervised learning Some reference material and a cool movie; https://arxiv.org/abs/1803.05316 category theory book https://math.mit.edu/~dspivak/teaching/sp18/ the class https://www.youtube.com/watch?v=nq6iPZVUxZU
https://arxiv.org/pdf/2008.02217.pdf - Hopfield Networks is All You Need
https://arxiv.org/pdf/1912.02762.pdf - Normalizing Flows for Probabilistic Modeling and Inference
https://arxiv.org/pdf/2007.02168.pdf - Scalable Differentiable Physics for Learning and Control
https://arxiv.org/pdf/1903.11239v3 Tossingbot
https://arxiv.org/pdf/2002.05709.pdf - A Simple Framework for Contrastive Learning of Visual Representations
https://arxiv.org/pdf/2002.11089.pdf - Rewriting History with Inverse RL: Hindsight Inference for Policy Improvement
https://arxiv.org/pdf/2002.11089.pdf - Rewriting History with Inverse RL: Hindsight Inference for Policy Improvement
https://www.osapublishing.org/DirectPDFAccess/C6D6B2C3-953C-4461-695B6E5E2F993943_415059/prj-7-8-823.pdf?da=1&id=415059&seq=0&mobile=no --Nanophotonic media for artificial neural inference
https://arxiv.org/pdf/1910.02789.pdf - Language is Power: Representing States Using Natural Language in Reinforcement Learning
https://deepmind.com/blog/article/AlphaFold-Using-AI-for-scientific-discovery - Protein folding paper.
https://arxiv.org/abs/2001.04451 Reformer, the efficient transformer
https://ai.googleblog.com/2020/01/reformer-efficient-transformer.html
https://arxiv.org/pdf/1906.05717.pdf - Unsupervised Monocular Depth and Ego-motion Learning with Structure and Semantics
https://arxiv.org/pdf/1912.09524.pdf - Evolving ab initio trading strategies in heterogeneous environments
https://arxiv.org/pdf/1911.05892.pdf - Reinforcement Learning for Market Making in Multi-agent Dealer Market
https://www.nature.com/articles/s41586-019-1724-z.epdf?author_access_token=lZH3nqPYtWJXfDA10W0CNNRgN0jAjWel9jnR3ZoTv0PSZcPzJFGNAZhOlk4deBCKzKm70KfinloafEF1bCCXL6IIHHgKaDkaTkBcTEv7aT-wqDoG1VeO9-wO3GEoAMF9bAOt7mJ0RWQnRVMbyfgH9A%3D%3D
https://www.gwern.net/docs/rl/2019-vinyals.pdf
https://deepmind.com/blog/article/AlphaStar-Grandmaster-level-in-StarCraft-II-using-multi-agent-reinforcement-learning
https://arxiv.org/pdf/1911.04252.pdf - Self-training with Noisy Student improves ImageNet classification
https://arxiv.org/pdf/1910.12713.pdf - Few-shot video-video synthesis
https://arxiv.org/pdf/1906.11883.pdf - Unsupervised learning of Object Keypoints for Perception and Control
https://arxiv.org/pdf/1710.03748.pdf - Emergent Complexity via Multi-Agent Competition
https://openai.com/blog/competitive-self-play/
https://arxiv.org/pdf/1703.04908.pdf - Emergence of Grounded Compositional Language in Multi-Agent Populations
https://arxiv.org/pdf/1909.07528.pdf - Emergent tool use from multi agent autocurricula
https://openai.com/blog/emergent-tool-use/
https://arxiv.org/pdf/1901.00949.pdf - Machine Teaching in Hierarchical Genetic Reinforcement Learning: Curriculum Design of Reward Functions for Swarm Shepherding
https://arxiv.org/pdf/1812.01729.pdf - Boltzman Generators - Sampling equilibrium states of many body systems with deep learning
https://arxiv.org/pdf/1907.10599.pdf - Fine Grained Spectral Perspective on Neural Networks
https://arxiv.org/pdf/1906.08237.pdf - XLNet Generalized autoregressive pretraining for language understanding
https://arxiv.org/pdf/1905.09272.pdf - Data efficient image recognition with contrastive predictive coding.
https://arxiv.org/pdf/1904.10509.pdf - Generating long sequences with sparse transformers
https://arxiv.org/pdf/1807.03748.pdf - Representation learning with contrastive predictive coding.
https://arxiv.org/pdf/1906.08253.pdf - When to trust your model: model-based policy optimization
https://arxiv.org/pdf/1901.09321.pdf - Fixup initialization - residual learning without normalization
http://proceedings.mlr.press/v97/mahoney19a/mahoney19a.pdf - Traditional and heavy tailed self regularization in neural net models
https://arxiv.org/pdf/1804.08838.pdf - Measuring intrinsic dimension of objective landscapes
https://arxiv.org/abs/1810.09536 - Ordered Neurons: Integrating Tree Structures into Recurrent Neural Networks
https://arxiv.org/pdf/1812.05159.pdf - An empirical study of example forgetting during neural network training.
https://arxiv.org/pdf/1812.00417.pdf - Snorkel Drybell - A case study in weak supervision at industrial scale
https://arxiv.org/pdf/1905.04981.pdf - Modelling instance level annotator reliability for natural language labelling
https://arxiv.org/pdf/1901.09321.pdf - Fixup Initialization: Residual Learning without Normalization
https://d4mucfpksywv.cloudfront.net/better-language-models/language_models_are_unsupervised_multitask_learners.pdf - Language Models are Unsupervised Multitask Learners.
https://arxiv.org/pdf/1811.00995.pdf - Invertible Residual Networks
https://arxiv.org/pdf/1904.01681.pdf - Augmented Neural ODE's
https://arxiv.org/pdf/1901.00596.pdf - Comprehensive Survey of Graph Neural Nets
https://github.com/rusty1s/pytorch_geometric
https://arxiv.org/pdf/1901.00596.pdf - Comprehensive Survey of Graph Neural Nets
https://papers.nips.cc/paper/7539-optimal-algorithms-for-non-smooth-distributed-optimization-in-networks.pdf - nips award winner
https://papers.nips.cc/paper/8200-non-delusional-q-learning-and-value-iteration.pdf - Non-delusional Q-learning and Value Iteration
https://arxiv.org/pdf/1706.03762.pdf - attention is all you need - Vaswani
https://github.com/jadore801120/attention-is-all-you-need-pytorch - easier to read code
https://www.youtube.com/watch?v=S0KakHcj_rs
https://tdls.a-i.science/events/2018-10-22/
https://tdls.a-i.science/events/2019-02-04/
http://nlp.seas.harvard.edu/2018/04/03/attention.html
https://arxiv.org/pdf/1806.02643.pdf - Re-evalating Evaluation
https://arxiv.org/pdf/1812.11951.pdf - Learning to Design RNA
https://arxiv.org/pdf/1901.02860.pdf - Transformer XL - Attentive Language Models, Beyond a fixed length context
https://arxiv.org/pdf/1809.06646.pdf - Model Free Adaptive Optimal Control of Sequential Manufacturing Process Using Reinforcement Learning
https://arxiv.org/pdf/1806.07366.pdf - Neural Ordinary Differential Equations - Top paper NIPS2019
https://arxiv.org/pdf/1606.05312.pdf - Successor Features for Transfer in Reinforcement Learning
http://proceedings.mlr.press/v37/schaul15.pdf - Universal Value Function Approximators
http://proceedings.mlr.press/v80/barreto18a/barreto18a.pdf - Transfer in deep reinforcement learning using successor features and generalised policy improvement.
https://www.youtube.com/watch?v=YDCPHekLUI4&t=1053s - Tom Schaul
https://www.youtube.com/watch?v=OCHwXxSW70o - Tejas Kulkarni
https://arxiv.org/pdf/1812.07626.pdf - Universal Successor Features Approximators
https://arxiv.org/pdf/1810.12715.pdf - On the Effectiveness of Interval Bound Propagation for Training Verifiably Robust Models
https://openreview.net/pdf?id=S1x4ghC9tQ - Temporal Difference Variational Autoencoder
https://openreview.net/pdf?id=S1JHhv6TW - Boosting Dilated Convolution with Mixed Tensor Decompositions
https://arxiv.org/pdf/1712.01208.pdf - The case for learned index structures
https://arxiv.org/abs/1809.07402 - Generalization properties of nn - Socher
https://einstein.ai/research/blog/identifying-generalization-properties-in-neural-networks - blog for above paper
https://arxiv.org/pdf/1802.05983.pdf - Disentangling by Factorising
https://arxiv.org/pdf/1804.00104.pdf - Learning Disentangled Joint, Discrete and Continuous Representations
https://arxiv.org/pdf/1807.05520.pdf - Deep Clustering for Unsupervised Learning of Visual Features
https://github.com/1Konny/FactorVAE
https://github.com/paruby/FactorVAE
https://github.com/nicolasigor/FactorVAE
https://arxiv.org/pdf/1810.12894.pdf - Exploration by Random Network Distillation - OpenAI
https://arxiv.org/pdf/1810.04805.pdf - Pre-trainged bi directional transformers for language translation
https://arxiv.org/pdf/1801.02613.pdf - Characterizing Adversarial Examples using Local Intrinsic Dimensionality
https://arxiv.org/pdf/1808.06670.pdf - Learning Deep Representations by Mutual Estimation Estimation and Maximization - Hjelm, Bengio
https://arxiv.org/pdf/1802.04364.pdf - Junction Tree Variational Auto-Encoder for Molecular Graph Generation
http://snap.stanford.edu/proj/embeddings-www/files/nrltutorial-part2-gnns.pdf
https://arxiv.org/pdf/1808.06601.pdf - Video to video synthesis https://github.com/NVIDIA/vid2vid - code
https://arxiv.org/pdf/1807.03146.pdf - Discovery of 3d keypoints from 2d image
https://arxiv.org/abs/1709.02371 - PWC-Net: CNNs for Optical Flow Using Pyramid, Warping, and Cost Volume," by Deqing Sun et al. (CVPR 2018)
Phil Ferrier will present the paper and run though his code for us. Phil's code is on his github reop:
https://github.com/philferriere/tfoptflow
https://arxiv.org/pdf/1807.03247.pdf - Intriguing failure (and improvement) to CNN for determining rotations.
https://arxiv.org/pdf/1803.03324.pdf - Learning Deep Generative Models of Graphs
https://arxiv.org/abs/1709.10082 - Optimally decentralized multi-robot collision avoidance w reinforcement learning.
https://github.com/TensorSwarm/TensorSwarm - Andreas Pasternak code for above
https://s3-us-west-2.amazonaws.com/openai-assets/research-covers/learning-dexterity/learning-dexterity-paper.pdf -Robot doing single hand manipulations.
https://www.theverge.com/2018/7/30/17621112/openai-robot-dexterity-dactyl-artificial-intelligence
https://arxiv.org/pdf/1711.03953.pdf - Breaking the softmax bottleneck
https://arxiv.org/pdf/1805.10829.pdf - SigSoftMax: Reanalyzing the softmax bottleneck
https://severelytheoretical.wordpress.com/2018/06/08/the-softmax-bottleneck-is-a-special-case-of-a-more-general-phenomenon/
https://arxiv.org/pdf/1807.01281.pdf - Human level performance in first person multiplayer games with population reinforcement learning.
https://deepmind.com/blog/capture-the-flag/
https://www.youtube.com/watch?v=steioHoiEms
https://arxiv.org/abs/1711.09846v2
https://arxiv.org/pdf/1611.05397.pdf
https://arxiv.org/pdf/1803.10122.pdf - schmidhuber paper on RL
https://deepmind.com/research/publications/neural-scene-representation-and-rendering/ - Rendering 3d scene
https://arxiv.org/pdf/1707.06347.pdf - Proximal Optimization Policies
https://openreview.net/pdf?id=BJOFETxR- - Learning to represent programs with graphs
https://openreview.net/pdf?id=BkisuzWRW - Zero Shot Visual Imitation - Reinforcement Learning
https://openreview.net/forum?id=HkL7n1-0b - Wasserstein Auto Encoders - one of ICLR top papers.
https://openreview.net/pdf?id=Hy7fDog0b - Ambient GAN - Generative Models from Lossy Measurements - ICLR top paper
https://arstechnica.com/science/2018/05/ai-trained-to-navigate-develops-brain-like-location-tracking/ - Grid representations in rat brain
https://deepmind.com/documents/200/Banino_at_al_final.pdf --
https://www.nature.com/articles/s41586-018-0102-6 --
https://arxiv.org/pdf/1712.06567.pdf - Deep Neuroevolution: Genetic Algorithms are a Competitive Alternative for
Training Deep Neural Networks for Reinforcement Learning
https://arxiv.org/pdf/1712.06560.pdf - Improving Exploration in Evolution Strategies for Deep Reinforcement
Learning via a Population of Novelty-Seeking Agents
https://eng.uber.com/deep-neuroevolution/ - Uber engineering blog post
https://arxiv.org/pdf/1801.10130.pdf - spherical CNN
https://arxiv.org/pdf/1710.07313.pdf - Using machine learning to replicate chaotic attractors
http://www.bmp.ds.mpg.de/tl_files/bmp/preprints/Zimmermann_Parlitz_preprint.pdf - paper to be published in "chaos"
https://www.quantamagazine.org/machine-learnings-amazing-ability-to-predict-chaos-20180418/ - blog post
https://arxiv.org/pdf/1711.10925.pdf - Deep Image Prior
https://dmitryulyanov.github.io/deep_image_prior - git hub from authors
https://box.skoltech.ru/index.php/s/ib52BOoV58ztuPM
http://mlexplained.com/2018/01/18/paper-dissected-deep-image-prior-explained/
http://fortune.com/2018/04/24/nvidia-artificial-intelligence-images/ - Article w video showing photo editing use
Finish Fractal AI
https://arxiv.org/pdf/1711.07971.pdf - non-local filtering
http://lanl.arxiv.org/pdf/1803.05049v1 - Fractal AI
https://arxiv.org/pdf/1803.04831.pdf - IndRNN longer deeper RNN's
https://arxiv.org/pdf/1711.10433.pdf - parallel wavenet
https://arxiv.org/pdf/1708.04552.pdf - regularizing convnet with cutout (desert paper)
http://www.cs.toronto.edu/~jmartens/docs/Deep_HessianFree.pdf - will get short presentation on this one.
https://arxiv.org/pdf/1802.03268.pdf - Efficient Neural Architecture Search via Parameter Sharing
https://github.com/carpedm20/ENAS-pytorch
some related papers and reviews.
https://arxiv.org/pdf/1708.05344.pdf - One shot architecture search
https://openreview.net/forum?id=ByQZjx-0-
and
https://openreview.net/forum?id=rydeCEhs-
https://arxiv.org/abs/1703.10135 - tacotron - end-to-end speech synthesis
https://arxiv.org/pdf/1712.05884.pdf - tacotron 2
https://research.googleblog.com/2017/12/tacotron-2-generating-human-like-speech.html -
https://github.com/A-Jacobson/tacotron2 - pytorch code
http://research.baidu.com/deep-speech-3%EF%BC%9Aexploring-neural-transducers-end-end-speech-recognition/
https://arxiv.org/pdf/1705.09792.pdf - Deep Complex Networks
https://arxiv.org/pdf/1801.10308.pdf - Nested LSTM's
https://arxiv.org/pdf/1705.10142.pdf - KRU from Fair
https://github.com/hannw/nlstm - tf code for Nested LSTM
http://openaccess.thecvf.com/content_cvpr_2017/papers/Khoreva_Simple_Does_It_CVPR_2017_paper.pdf - Weakly Supervised Instance and Semantic Segmentation
https://www.mpi-inf.mpg.de/departments/computer-vision-and-multimodal-computing/research/weakly-supervised-learning/simple-does-it-weakly-supervised-instance-and-semantic-segmentation/
https://github.com/philferriere/tfwss - Phil Ferriere's code
https://drive.google.com/file/d/1wPHMA4PqygawvIxRiy-2ZMKcpUO447cz/view?usp=sharing - mehul's notebook on segmentation
https://arxiv.org/pdf/1511.06939.pdf - using rnn for recommendation system
https://static.googleusercontent.com/media/research.google.com/en//pubs/archive/46488.pdf - latest paper on rnn for recommendation
https://arxiv.org/pdf/1709.04511.pdf - Empirical study of multi-agent RL
https://github.com/geek-ai/1m-agents - code
https://arxiv.org/pdf/1704.00028.pdf - Improvements in Wasserstein GAN training
https://arxiv.org/pdf/1710.02298.pdf - Combining improvements in deep reinforcement learning
https://openreview.net/pdf?id=HJWLfGWRb - follow-on to capsule network paper
https://www.youtube.com/watch?v=pPN8d0E3900
https://www.youtube.com/watch?v=2Kawrd5szHE
https://github.com/ageron/handson-ml/blob/master/extra_capsnets.ipynb
https://github.com/naturomics/CapsNet-Tensorflow
https://medium.com/ai%C2%B3-theory-practice-business/understanding-hintons-capsule-networks-part-ii-how-capsules-work-153b6ade9f66
https://arxiv.org/pdf/1710.09829.pdf - Dynamic routing between capsules - Hinton
https://arxiv.org/pdf/1701.01724.pdf - DeepStack: Expert-Level Artificial Intelligence in Heads-Up No-Limit Poker
https://deepmind.com/documents/119/agz_unformatted_nature.pdf - alpha zero paper
https://webdocs.cs.ualberta.ca/~mmueller/talks/2016-LeeSedol-AlphaGo.pdf - some slides
https://arxiv.org/pdf/1703.10593.pdf - cycle consistent GANs
https://arxiv.org/pdf/1503.02406.pdf Naftali Tishby and Noga Zaslavsky. information bottleneck principle.
https://www.cs.huji.ac.il/labs/learning/Papers/allerton.pdf - Naftali Tishby, Fernando C. Pereira, and William Bialek. The information bottleneck method.
Mask R-CNN
https://arxiv.org/abs/1703.06870
And these are prerequisites (read at least Fast R-CNN and Faster R-CNN)
R-CNN
https://arxiv.org/abs/1311.2524
Fast R-CNN
https://arxiv.org/pdf/1504.08083.pdf
Faster R-CNN
https://arxiv.org/abs/1506.01497 Feature Pyramid Networks
https://arxiv.org/abs/1612.03144
https://arxiv.org/pdf/1703.00810.pdf - Opening the Black Box of Neural Nets via Information
https://www.youtube.com/watch?v=ekUWO_pI2M8
https://www.youtube.com/watch?v=bLqJHjXihK8
https://arxiv.org/pdf/1501.00092.pdf - super resolution first paper
https://arxiv.org/abs/1608.00367 - super resolution second paper
https://arxiv.org/abs/1604.03901 - Single-Image Depth Perception in the Wild
https://arxiv.org/pdf/1706.08947.pdf - Exploring generalization in deep networks.
https://arxiv.org/pdf/1705.02550.pdf - nvidia drone nav
https://github.com/NVIDIA-Jetson/redtail/wiki - code
http://citeseerx.ist.psu.edu/viewdoc/download?doi=10.1.1.365.5060&rep=rep1&type=pdf - hyperneat ref
https://arxiv.org/pdf/1609.09106.pdf - Hypernet ref
http://blog.otoro.net/2016/09/28/hyper-networks/ - blog on hypernet
https://www.youtube.com/watch?v=-8oyTYViuJ4 - vid on hyperNeat
http://eplex.cs.ucf.edu/hyperNEATpage/HyperNEAT.html - blog on hyperNeat
https://arxiv.org/pdf/1708.05344.pdf - SMASH: One-Shot Model Architecture Search through HyperNetworks https://www.youtube.com/watch?v=79tmPL9AL48 - youtube vid on SMASH
https://arxiv.org/pdf/1706.02515.pdf - Self Normalizing Neural Networks - Hochreiter
https://arxiv.org/pdf/1606.01541.pdf - Reinforcement Learning for Dialog Generation - Jurafsky
https://github.com/liuyuemaicha/Deep-Reinforcement-Learning-for-Dialogue-Generation-in-tensorflow - tensorflow code for same
https://github.com/jiweil/ - some related code
https://arxiv.org/pdf/1612.00563.pdf - self critical training for image captioning - RL for text prob.
Some papers referenced by Jurafsky paper
[1506.05869] A Neural Conversational Model - Vinyals and Le
https://arxiv.org/abs/1604.04562 - Dialogue generation system - Wen
https://arxiv.org/pdf/1705.04304.pdf - A Deep Reinforced Model for Abstractive Summarization - socher
https://arxiv.org/pdf/1706.01433.pdf - visual interaction networks - deep mind
https://arxiv.org/pdf/1706.01427.pdf - neural model for relational reasoning - deep mind
Guest Speaker - Using FPGA to speed CNN.
https://arxiv.org/pdf/1703.03130.pdf - A structured self-attentive sentence embedding - Lin and Bengio
https://github.com/dennybritz/deeplearning-papernotes/blob/master/notes/self_attention_embedding.md (review)
https://github.com/yufengm/SelfAttentive code
https://github.com/Diego999/SelfSent code
https://arxiv.org/pdf/1706.03762.pdf - attention is all you need - Vaswani
https://github.com/tensorflow/tensor2tensor/tree/master/tensor2tensor/models
https://github.com/jadore801120/attention-is-all-you-need-pytorch - easier to read code
https://arxiv.org/pdf/1607.06450.pdf - layer normalization paper - hinton
https://www.youtube.com/watch?v=nR74lBO5M3s - google translate paper - youtube video
https://arxiv.org/pdf/1609.08144.pdf - google translate paper -
https://arxiv.org/pdf/1706.03762.pdf - attention is all you need - Vaswani
https://github.com/tensorflow/tensor2tensor/tree/master/tensor2tensor/models
https://github.com/jadore801120/attention-is-all-you-need-pytorch - easier to read code
https://arxiv.org/pdf/1607.06450.pdf - layer normalization paper - hinton
http://www.machinelearning.org/proceedings/icml2006/047_Connectionist_Tempor.pdf - A. Graves, S. Fernandez, F. Gomez, and J. Schmidhuber
https://www.reddit.com/r/MachineLearning/comments/6jdi87/r_question_about_positional_encodings_used_in/
https://arxiv.org/pdf/1705.03122.pdf - convolutional sequence to sequence learning
https://arxiv.org/pdf/1706.03762.pdf - attention is all you need - Vaswani
http://www.machinelearning.org/proceedings/icml2006/047_Connectionist_Tempor.pdf - A. Graves, S. Fernandez, F. Gomez, and J. Schmidhuber
https://arxiv.org/pdf/1701.02720.pdf - RNN for end to end voice recognition
New reinforcement learning results -- Too cool for school. Watch the video and you'll be hooked.
https://www.youtube.com/watch?v=2vnLBb18MuQ&feature=em-subs_digest
http://www.cs.ubc.ca/~van/papers/2017-TOG-deepLoco/index.html - paper
https://www.microsoft.com/en-us/research/wp-content/uploads/2016/02/HintonDengYuEtAl-SPM2012.pdf - comparison of RNN and HMM for speech recognition
https://arxiv.org/pdf/1412.6572.pdf - Explaining and Harnessing Adversarial Examples
https://arxiv.org/abs/1704.03453 - The Space of Transferable Adversarial Examples
https://discourse-production.oss-cn-shanghai.aliyuncs.com/original/3X/1/5/15ba4cef726cab390faa180eb30fd82b693469f9.pdf - Using TPU for data center
Reservoir Computing by Felix Grezes. http://www.gc.cuny.edu/CUNY_GC/media/Computer-Science/Student%20Presentations/Felix%20Grezes/Second_Exam_Survey_Felix_Grezes_9_04_2014.pdf
Slides by Felix Grezes: Reservoir Computing for Neural Networks
http://www.gc.cuny.edu/CUNY_GC/media/Computer-Science/Student%20Presentations/Felix%20Grezes/Second_Exam_Slides_Felix_Grezes_9-14-2014.pdf
(more at: http://speech.cs.qc.cuny.edu/~felix/ )
This is a short, very useful backgrounder on randomized projections,
here used for compressed sensing, in a blog post by Terence Tao
https://terrytao.wordpress.com/2007/04/13/compressed-sensing-and-single-pixel-cameras/
and the same story told with illustrations on the Nuit Blanche blog:
http://nuit-blanche.blogspot.com/2007/07/how-does-rice-one-pixel-camera-work.html
(BTW http://nuit-blanche.blogspot.com is a tremendous website.)
If we have time, we may discuss this paper:
Information Processing Using a Single Dynamical Node as Complex System.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3195233/pdf/ncomms1476.pdf
https://arxiv.org/pdf/1603.08678.pdf - Instance-sensitive Fully Convolutional Networks
https://arxiv.org/pdf/1611.07709.pdf - Fully Convolutional Instance-aware Semantic Segmentation
https://arxiv.org/pdf/1703.03864.pdf - Sutskever paper on using evolutionary systems for optimizing RL prob
http://jmlr.csail.mit.edu/papers/volume15/wierstra14a/wierstra14a.pdf - ES paper with algo used in Sutskever paper
Aurobindo Tripathy will reprise a talk he's going to give at Embedded Summit this year. His talk will survey recent progress in object detection from RCNN to Single Shot MultiBox Detector and Yolo 9000.
https://arxiv.org/pdf/1612.05424.pdf - Unsupervised Pixel-level domain adaptation with generative adversarial networks
https://arxiv.org/pdf/1701.06547.pdf - adversarial learning for neural dialog generation
https://arxiv.org/pdf/1612.02699.pdf - Deep Supervision with Shape Concepts for Occlusion-Aware 3D Object Parsing
Zeeshan's slides are in the folder with his name on it. Along with his descriptions of his own ground-breaking work, he gives an excellent history of efforts to identify 3d objects from 2d images.
https://arxiv.org/pdf/1506.07285.pdf - Ask me anything - Socher
https://github.com/YerevaNN/Dynamic-memory-networks-in-Theano - Code and implementation notes.
https://www.youtube.com/watch?v=FCtpHt6JEI8&t=27s - Socher presentation of material
https://arxiv.org/pdf/1701.06538v1.pdf - Outrageously large neural networks
https://arxiv.org/pdf/1505.00387v2.pdf - Highway networks
https://arxiv.org/pdf/1507.06228.pdf - Also highway networks - different examples
https://arxiv.org/pdf/1607.03474v3.pdf - Recurrent Highway Networks
https://arxiv.org/pdf/1603.03116v2.pdf - Low-rank pass-through RNN's follow-on to unitary rnn https://github.com/Avmb/lowrank-gru - theano code
https://arxiv.org/abs/1612.03242 - Stack Gan Paper
https://github.com/hanzhanggit/StackGAN - Code
https://arxiv.org/pdf/1511.06464v4.pdf - Unitary Evolution RNN https://github.com/amarshah/complex_RNN - theano code
Cheuksan Edward Wang Talk
https://arxiv.org/pdf/1612.04642v1.pdf - rotation invariant cnn
https://github.com/deworrall92/harmonicConvolutions - tf code for harmonic cnn
http://visual.cs.ucl.ac.uk/pubs/harmonicNets/index.html - blog post by authors
https://arxiv.org/pdf/1602.02218v2.pdf - using typing to improve RNN behavior
http://jmlr.org/proceedings/papers/v37/jozefowicz15.pdf - exploration of alternative LSTM architectures
https://arxiv.org/pdf/1611.01576.pdf - Socher qRnn paper
https://arxiv.org/pdf/1604.02135v2.pdf - latest segmentation fair
https://github.com/MarvinTeichmann/tensorflow-fcn - code for segmenter
https://arxiv.org/pdf/1506.06204.pdf - Object segmentation
https://arxiv.org/pdf/1603.08695v2.pdf - refinement of above segmentation paper
https://code.facebook.com/posts/561187904071636/segmenting-and-refining-images-with-sharpmask/ - blog post
https://github.com/facebookresearch/deepmask - torch code for deepmask
https://arxiv.org/pdf/1506.01497v3.pdf
people.eecs.berkeley.edu/~rbg/slides/rbg-defense-slides.pdf - Girshick thesis slides
Check edge boxes and selective search
https://arxiv.org/pdf/1406.4729v4.pdf - key part of architecture
https://github.com/smallcorgi/Faster-RCNN_TF - excellent code
https://people.eecs.berkeley.edu/~rbg/papers/r-cnn-cvpr.pdf - RCNN
https://arxiv.org/pdf/1504.08083v2.pdf - RCNN - first in series
https://arxiv.org/pdf/1506.01497v3.pdf - Faster R-CNN
http://techtalks.tv/talks/rich-feature-hierarchies-for-accurate-object-detection-and-semantic-segmentation/60254/ - video of Girshick talk
https://arxiv.org/pdf/1506.02025v3.pdf - Spatial transformer networks
https://github.com/daviddao/spatial-transformer-tensorflow - tf code for above
https://github.com/jazzsaxmafia/show_attend_and_tell.tensorflow - tf code for attention-captioning http://cs.stanford.edu/people/karpathy/densecap/ - karpathy captioning https://arxiv.org/pdf/1412.2306v2.pdf - earlier karpathy captioning paper
https://webdocs.cs.ualberta.ca/~sutton/book/the-book.html - Deep dive into reinforcement learning - Sutton and Barto - Chapters 1 and 2.
https://arxiv.org/pdf/1608.06993v1.pdf - DenseNet. New reigning champion image classifier
https://github.com/liuzhuang13/DenseNet - lua code
The DenseNet paper is straight-forward, so we're also going to start on image captioning
http://www.cs.toronto.edu/~zemel/documents/captionAttn.pdf
http://kelvinxu.github.io/projects/capgen.html
http://people.ee.duke.edu/~lcarin/Yunchen9.25.2015.pdf - slides for caption attention
collections of captioning papers.
https://github.com/kjw0612/awesome-deep-vision#image-captioning - images
https://github.com/kjw0612/awesome-deep-vision#video-captioning - video
http://www.mit.edu/~dimitrib/NDP_Encycl.pdf - (early) Bersekas paper on RL, policy and value iteration
http://www.nervanasys.com/demystifying-deep-reinforcement-learning/?imm_mid=0e2d7e&cmp=em-data-na-na-newsltr_20160420 - blog post on RL. Nice coverage of value iteration
https://github.com/carpedm20/pixel-rnn-tensorflow - tensorflow code for pixel rnn (and cnn)
https://arxiv.org/pdf/1606.05328v2.pdf - Conditional Image Generation with PixelCNN decoders
https://arxiv.org/pdf/1601.06759v3.pdf - Pixel RNN
https://drive.google.com/file/d/0B3cxcnOkPx9AeWpLVXhkTDJINDQ/view - wavenet Generative Audio
https://deepmind.com/blog/wavenet-generative-model-raw-audio/ - wavenet blog
http://www.gitxiv.com/posts/fepYG4STYaej3KSPZ/densely-connected-convolutional-netowork-densenet
http://arxiv.org/pdf/1410.3916v11.pdf - original memory networks
https://arxiv.org/pdf/1606.03126v1.pdf - key/value memory augmented nn
http://www.thespermwhale.com/jaseweston/icml2016/icml2016-memnn-tutorial.pdf#page=87 - tutorial on memory networks in language understanding
https://arxiv.org/pdf/1410.5401v2.pdf - Neural Turing Machines
https://github.com/carpedm20/NTM-tensorflow
https://www.youtube.com/watch?v=_H0i0IhEO2g - Alex Graves presentation at microsoft research
http://www.robots.ox.ac.uk/~tvg/publications/talks/NeuralTuringMachines.pdf - slides for ntm
http://arxiv.org/pdf/1410.3916v11.pdf - original memory networks
https://arxiv.org/pdf/1606.03126v1.pdf - key/value memory augmented nn
http://www.thespermwhale.com/jaseweston/icml2016/icml2016-memnn-tutorial.pdf#page=87 - tutorial on memory networks in language understanding
https://arxiv.org/pdf/1605.07648v1.pdf - fractal net - alternative to resnet for ultra-deep convolution
https://github.com/edgelord/FractalNet - tf code
http://www.gitxiv.com/posts/ibA8QEu8bvBJSDxr9/fractalnet-ultra-deep-neural-networks-without-residuals
https://arxiv.org/pdf/1602.01783v2.pdf - new RL architecture - deep mind
Code:
https://github.com/Zeta36/Asynchronous-Methods-for-Deep-Reinforcement-Learning - tf
https://github.com/miyosuda/async_deep_reinforce - tf
https://github.com/coreylynch/async-rl - keras (tf)
https://github.com/muupan/async-rl - chainer (good discussion)
https://arxiv.org/pdf/1607.02533v1.pdf - Hardening deep networks to adversarial examples.
http://www.gitxiv.com/posts/HQJ3F9YzsQZ3eJjpZ/model-free-episodic-control - deep mind gitxiv paper and code on github https://github.com/sudeepraja/Model-Free-Episodic-Control - other code https://github.com/ShibiHe/Model-Free-Episodic-Control
https://arxiv.org/pdf/1406.2661.pdf - originating paper on generative adversarial net (gan) - goodfellow, bengio
http://arxiv.org/pdf/1511.06434v2.pdf - deep cnn gan - radford
https://github.com/Newmu/dcgan_code - theano code for cnn gan - radford
http://www.gitxiv.com/posts/HQJ3F9YzsQZ3eJjpZ/model-free-episodic-control - deep mind gitxiv paper and code on github
Papers -
https://drive.google.com/file/d/0B8Dg3PBX90KNWG5KQXNQOFlBLU1JWWVONkN1UFpnbUR6Y0cw/view?pref=2&pli=1 - Using Stochastic RNN for temporal anomaly detection
https://home.zhaw.ch/~dueo/bbs/files/vae.pdf - cover math
https://arxiv.org/pdf/1401.4082v3.pdf - Rezende - Other Original VAE paper
Code Review -
https://github.com/oduerr/dl_tutorial/blob/master/tensorflow/vae/vae_demo.ipynb
https://github.com/oduerr/dl_tutorial/blob/master/tensorflow/vae/vae_demo-2D.ipynb
Papers:
http://arxiv.org/pdf/1410.5401v2.pdf - Neural Turing Machines - Graves et. al.
https://arxiv.org/pdf/1605.06065v1.pdf - One Shot Learning - DeepMind
Code:
http://icml.cc/2016/reviews/839.txt
https://github.com/brendenlake/omniglot
https://github.com/tristandeleu/ntm-one-shot
https://github.com/MLWave/extremely-simple-one-shot-learning
Papers - Using VAE for anomaly detection
https://arxiv.org/pdf/1411.7610.pdf - Stochastic Recurrent Networks
https://drive.google.com/file/d/0B8Dg3PBX90KNWG5KQXNQOFlBLU1JWWVONkN1UFpnbUR6Y0cw/view?pref=2&pli=1 - Using Stochastic RNN for temporal anomaly detection
Papers to read:
http://www.thespermwhale.com/jaseweston/ram/papers/paper_16.pdf
http://snowedin.net/tmp/Hochreiter2001.pdf -
Comments / Code
http://icml.cc/2016/reviews/839.txt
https://github.com/brendenlake/omniglot
https://github.com/tristandeleu/ntm-one-shot
https://github.com/MLWave/extremely-simple-one-shot-learning
https://www.periscope.tv/hugo_larochelle/1ypJdnPRYEoKW
Papers to read:
http://arxiv.org/pdf/1312.6114v10.pdf - variational autoencoders - U of Amsterdam - Kingma and Welling
http://arxiv.org/pdf/1310.8499v2.pdf - deep autoregressive networks - deep mind
https://arxiv.org/abs/1606.05908 - tutorial on vae
Commentaries/Code
https://jmetzen.github.io/2015-11-27/vae.html - metzen - code and discussion
http://blog.keras.io/building-autoencoders-in-keras.html - chollet - discusses different autoencoders, gives keras code.
Recurrent network for image generation - Deep Mind
https://arxiv.org/pdf/1502.04623v2.pdf
Background and some references cited
http://blog.evjang.com/2016/06/understanding-and-implementing.html - blog w. code for VAE
http://arxiv.org/pdf/1312.6114v10.pdf - Variational Auto Encoder
https://jmetzen.github.io/2015-11-27/vae.html - tf code for variational auto-encoder
https://www.youtube.com/watch?v=P78QYjWh5sM
https://arxiv.org/pdf/1401.4082.pdf - stochastic backpropagation and approx inference - deep mind
http://www.cs.toronto.edu/~fritz/absps/colt93.html - keep neural simple by minimizing descr length - hinton
https://github.com/vivanov879/draw - code
Recurrent models of visual attention - Deep Mind
https://papers.nips.cc/paper/5542-recurrent-models-of-visual-attention.pdf
http://arxiv.org/pdf/1410.5401v2.pdf - Neural Turing Machines - Graves et. al.
https://arxiv.org/pdf/1605.06065v1.pdf - One Shot Learning - DeepMind
http://www.shortscience.org/paper?bibtexKey=journals/corr/1605.06065 - Larochell comments on One-Shot paper
https://github.com/shawntan/neural-turing-machines - Code
https://www.reddit.com/r/MachineLearning/comments/2xcyrl/i_am_j%C3%BCrgen_schmidhuber_ama/cp4ecce - schmidhuber's comments
http://www.thespermwhale.com/jaseweston/ram/papers/paper_16.pdf
http://snowedin.net/tmp/Hochreiter2001.pdf -
Reviews:
http://icml.cc/2016/reviews/839.txt
Code
https://github.com/brendenlake/omniglot
https://github.com/tristandeleu/ntm-one-shot
https://github.com/MLWave/extremely-simple-one-shot-learning
Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learning:
http://arxiv.org/pdf/1602.07261v1.pdf
Visualizing and Understanding RNN:
https://arxiv.org/pdf/1506.02078v2.pdf
Google inception paper - origin of 1x1 convolution layers
http://arxiv.org/pdf/1409.4842v1.pdf
Image segmentation with deep encoder-decoder https://arxiv.org/pdf/1511.00561.pdf
Compressed networks, reducing flops by pruning https://arxiv.org/pdf/1510.00149.pdf http://arxiv.org/pdf/1602.07360v3.pdf
Word2Vec meets LDA: http://arxiv.org/pdf/1605.02019v1.pdf - Paper
https://twitter.com/chrisemoody - Chris Moody's twitter with links to slides etc.
http://qpleple.com/topic-coherence-to-evaluate-topic-models/ - writeup on topic coherence
https://arxiv.org/pdf/1603.05027v2.pdf - Update on microsoft resnet - identity mapping
http://gitxiv.com/posts/MwSDm6A4wPG7TcuPZ/recurrent-batch-normalization - batch normalization w. RNN
Go playing DQN - AlphaGo https://gogameguru.com/i/2016/03/deepmind-mastering-go.pdf https://m.youtube.com/watch?sns=em&v=pgX4JSv4J70 - video of slide presentation on paper https://en.m.wikipedia.org/wiki/List_of_Go_games#Lee.27s_Broken_Ladder_Game - Handling "ladders" in alphgo https://en.m.wikipedia.org/wiki/Ladder_(Go) - ladders in go
Deep Residual Learning for Image Recognition
http://arxiv.org/pdf/1512.03385v1.pdf
References: http://arxiv.org/pdf/1603.05027v2.pdf - Identity mapping paper
Code:
https://keunwoochoi.wordpress.com/2016/03/09/residual-networks-implementation-on-keras/ - keras code
https://github.com/ry/tensorflow-resnet/blob/master/resnet.py - tensorflow code
https://github.com/tensorflow/tensorflow/blob/master/tensorflow/examples/skflow/resnet.py
Playing Atari with Deep Reinforcement Learning
https://www.cs.toronto.edu/~vmnih/docs/dqn.pdf
http://gitxiv.com/posts/MwSDm6A4wPG7TcuPZ/recurrent-batch-normalization - Batch Normalization for RNN
Playing Atari with Deep Reinforcement Learning
https://www.cs.toronto.edu/~vmnih/docs/dqn.pdf
Related references:
This adds 'soft' and 'hard' attention and the 4 frames are replaced with an LSTM layer:
http://gitxiv.com/posts/NDepNSCBJtngkbAW6/deep-attention-recurrent-q-network
http://home.uchicago.edu/~arij/journalclub/papers/2015_Mnih_et_al.pdf - Nature Paper
http://www.nature.com/nature/journal/v518/n7540/full/nature14236.html - videos at the bottom of the page
http://llcao.net/cu-deeplearning15/presentation/DeepMindNature-preso-w-David-Silver-RL.pdf - David Silver's slides
http://www.cogsci.ucsd.edu/~ajyu/Teaching/Cogs118A_wi09/Class0226/dayan_watkins.pdf
http://www0.cs.ucl.ac.uk/staff/d.silver/web/Teaching.html - David Silver
Implementation Examples:
http://stackoverflow.com/questions/35394446/why-doesnt-my-deep-q-network-master-a-simple-gridworld-tensorflow-how-to-ev?rq=1
http://www.danielslater.net/2016/03/deep-q-learning-pong-with-tensorflow.html
Gated Feedback Recurrent Neural Networks
https://arxiv.org/pdf/1502.02367v4.pdf)
-Background Material
http://arxiv.org/pdf/1506.00019v4.pdf - Lipton's excellent review of RNN
http://www.nehalemlabs.net/prototype/blog/2013/10/10/implementing-a-recurrent-neural-network-in-python/ - Discussion of RNN and theano code for Elman network - Tiago Ramalho
http://deeplearning.cs.cmu.edu/pdfs/Hochreiter97_lstm.pdf - Hochreiter's original paper on LSTM
https://www.youtube.com/watch?v=izGl1YSH_JA - Hinton video on LSTM
-Skylar Payne's GF RNN code
https://github.com/skylarbpayne/hdDeepLearningStudy/tree/master/tensorflow
-Slides
https://docs.google.com/presentation/d/1d2keyJxRlDcD1LTl_zjS3i45xDIh2-QvPWU3Te29TuM/edit?usp=sharing
https://github.com/eadsjr/GFRNNs-nest/tree/master/diagrams/diagrams_formula
http://www.computervisionblog.com/2016/06/deep-learning-trends-iclr-2016.html
https://indico.io/blog/iclr-2016-takeaways/