Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.
Awesome Lists | Featured Topics | Projects
https://github.com/open-thought/system-2-research

System 2 Reasoning Link Collection
https://github.com/open-thought/system-2-research
Last synced: 2 months ago
JSON representation
System 2 Reasoning Link Collection
Host: GitHub
URL: https://github.com/open-thought/system-2-research
Owner: open-thought
License: apache-2.0
Created: 2024-08-03T10:15:46.000Z (3 months ago)
Default Branch: main
Last Pushed: 2024-08-30T09:32:58.000Z (3 months ago)
Last Synced: 2024-08-30T10:18:35.962Z (3 months ago)
Size: 79.1 KB
Stars: 379
Watchers: 15
Forks: 35
Open Issues: 2
Metadata Files:
- Readme: README.md
- License: LICENSE
Awesome Lists containing this project

README

        # OpenThought - System 2 Research Links

Here you find a collection of material (books, papers, blog-posts etc.) related to reasoning and cognition in AI systems. Specifically we want to cover agents, cognitive architectures, general problem solving strategies and self-improvement.

The term "System 2" in the page title refers to the slower, more deliberative, and more logical mode of thought as described by Daniel Kahneman in his book [Thinking, Fast and Slow](https://en.wikipedia.org/wiki/Thinking,_Fast_and_Slow).

You know a great resource we should add? Please see [How to contribute](#how-to-contribute).

### Weekly Meeting

We meet every Tuesday at 6 PM UTC in `#voice-room-1` of [Yannic Kilcher's discord server](https://ykilcher.com/discord). You can also reach us there any time asynchronously in the `#open-thought` channel.

## Cognitive Architectures

(looking for additional links & articles and summaries)

- [SOAR](https://en.wikipedia.org/wiki/Soar_(cognitive_architecture)) (State, Operator, And Result) by John Laird, Allen Newell, and Paul Rosenbloom

- [ACT-R](https://en.wikipedia.org/wiki/ACT-R) (Adaptive Control of Thought-Rational) by John Anderson at CMU

- [SPAUN](https://www.nengo.ai/nengo-spa/user-guide/spa-intro.html) (Semantic Pointer Architecture Unified Network) by Chris Eliasmith at Waterloo, [SPAUN 2.0](https://core.ac.uk/download/pdf/158325694.pdf) by Feng-Xuan Choo

- [ART](https://en.wikipedia.org/wiki/Adaptive_resonance_theory) (Adaptive resonance theory) by Stephen Grossberg and Gail Carpenter

- [CLARION](https://en.wikipedia.org/wiki/CLARION_(cognitive_architecture)) (Connectionist Learning with Adaptive Rule Induction ON-line) by Ron Sun

- [EPIC](https://web.eecs.umich.edu/~kieras/epic.html) (Executive Process/Interactive Control) by David Kieras and David Meyer

- [LIDA](https://en.wikipedia.org/wiki/LIDA_(cognitive_architecture)) (Learning Intelligent Distribution Agent) by Stan Franklin, [2015 Paper](https://digitalcommons.memphis.edu/cgi/viewcontent.cgi?article=1023&context=ccrg_papers)

- [Sigma](https://ict.usc.edu/pubs/The%20Sigma%20cognitive%20architecture%20and%20system.pdf) by Paul Rosenbloom

- [OpenCog](https://opencog.org/) by Ben Goertzel

- [NARS](https://cis.temple.edu/~pwang/NARS-Intro.html) (Non-Axiomatic Reasoning System) by Pei Wang

- [Icarus](http://www.isle.org/~langley/papers/icarus.csr17.pdf) by Pat Langley

- [MicroPsi](http://cognitive-ai.com/publications/assets/MicroPsiArchitectureICCM03.pdf) by Joscha Bach

- [Thousand Brains Theory](https://www.numenta.com/blog/2019/01/16/the-thousand-brains-theory-of-intelligence/) & [HTM](https://www.numenta.com/resources/htm/) (Hierarchical Temporal Memory) by Jeff Hawkins

- [SPH](https://ogma.ai/sph-technology-description/) (Sparse Predictive Hierarchie) by Eric Laukien

- [Leabra](https://github.com/emer/leabra) (Local, Error-driven and Associative, Biologically Realistic Algorithm), [2016 Paper](https://ccnlab.org/papers/OReillyHazyHerd16.pdf) by [Randall O'Reilly](https://ccnlab.org/people/oreilly/)

- [CogNGen](https://arxiv.org/abs/2310.15177) (COGnitive Neural GENerative system) by Alexander Ororbia and Mary Alexandria Kelly, see also [here](https://osf.io/preprints/osf/cew42) and [here](https://arxiv.org/abs/2204.00619)

- [KIX](https://arxiv.org/abs/2402.05346) (KIX: A Metacognitive Generalization Framework) by A. Kumar and Paul Schrater

- [ACE](https://arxiv.org/abs/2310.06775) (Autonomous Cognitive Entity) by David Shapiro et al., gh: [daveshap/ACE_Framework](https://github.com/daveshap/ACE_Framework)

- [Iterative Updating of Working Memory](https://arxiv.org/abs/2203.17255) by Jared Reser, [website](https://aithought.com/), [Video](https://youtu.be/R2H2Pl0I6EA?si=DlO0j-WxhG5TaJeN)

## Agent Papers

### LLM Based

- Self-Taught Reasoner: [STaR: Bootstrapping Reasoning With Reasoning](https://arxiv.org/abs/2203.14465)

- [Quiet-STaR: Language Models Can Teach Themselves to Think Before Speaking](https://arxiv.org/abs/2403.09629)

- [ReAct: Synergizing Reasoning and Acting in Language Models](https://arxiv.org/abs/2210.03629)

- [SWE-agent: Agent-Computer Interfaces Enable Automated Software Engineering](https://arxiv.org/abs/2405.15793)

- [The AI Scientist: Towards Fully Automated Open-Ended Scientific Discovery](https://arxiv.org/abs/2408.06292), gh: [SakanaAI/AI-Scientist](https://github.com/SakanaAI/AI-Scientist)

- [OpenDevin: An Open Platform for AI Software Developers as Generalist Agents](https://arxiv.org/abs/2407.16741)

- [TextGrad: Automatic "Differentiation" via Text](https://arxiv.org/abs/2406.07496)

- [Trace is the New AutoDiff -- Unlocking Efficient Optimization of Computational Workflows](https://arxiv.org/abs/2406.16218)

- [Agentless: Demystifying LLM-based Software Engineering Agents](https://arxiv.org/abs/2407.01489)

- [Competition-Level Code Generation with AlphaCode](https://arxiv.org/abs/2203.07814)

- [AI Agents That Matter](https://arxiv.org/pdf/2407.01502)

- [Sibyl: Simple yet Effective Agent Framework for Complex Real-world Reasoning](https://arxiv.org/abs/2407.10718)

- [Trial and Error: Exploration-Based Trajectory Optimization for LLM Agents](https://arxiv.org/abs/2403.02502)

- [Self-Rewarding Language Models](https://arxiv.org/abs/2401.10020)

- [ArchCode: Incorporating Software Requirements in Code Generation with Large Language Models](https://arxiv.org/abs/2408.00994)

- Self-Notes: [Learning to Reason and Memorize with Self-Notes](https://arxiv.org/abs/2305.00833)

- MedAgent-Zero: [Agent Hospital: A Simulacrum of Hospital with Evolvable Medical Agents](https://arxiv.org/abs/2405.02957)

- [Self-Discover: Large Language Models Self-Compose Reasoning Structures](https://arxiv.org/abs/2402.03620)

- [Cognitive Architectures for Language Agents](https://arxiv.org/abs/2309.02427)

- [Large Language Models Can Self-Improve At Web Agent Tasks](https://arxiv.org/abs/2405.20309)

- [AgentGen: Enhancing Planning Abilities for Large Language Model based Agent via Environment and Task Generation](https://arxiv.org/abs/2408.00764)

- [A Prefrontal Cortex-inspired Architecture for Planning in Large Language Models](https://arxiv.org/abs/2310.00194)

- [CLIN: A Continually Learning Language Agent for Rapid Task Adaptation and Generalization](https://arxiv.org/abs/2310.10134)

- [DeepSeek-Prover-V1.5: Harnessing Proof Assistant Feedback for Reinforcement Learning and Monte-Carlo Tree Search](https://arxiv.org/abs/2408.08152)

- GPT-Swarm [Language Agents as Optimizable Graphs](https://arxiv.org/abs/2402.16823)

- Survey: [Reasoning with Large Language Models, a Survey](https://arxiv.org/abs/2407.11511) (Jul 2024)

- Survey: [From LLMs to LLM-based Agents for Software Engineering: A Survey of Current, Challenges and Future](https://arxiv.org/abs/2408.02479) (Aug 2024)

- ADAS: [Automated Design of Agentic Systems](https://arxiv.org/abs/2408.08435)

- [Agent Q: Advanced Reasoning and Learning for Autonomous AI Agents](https://arxiv.org/abs/2408.07199)

- rStar: [Mutual Reasoning Makes Smaller LLMs Stronger Problem-Solvers](https://arxiv.org/abs/2408.06195)

- [IDEA: Enhancing the rule learning ability of language agent through Induction, DEuction, and Abduction](https://arxiv.org/abs/2408.10455)

- LAW: [Language Models, Agent Models, and World Models: The LAW for Machine Reasoning and Planning](https://arxiv.org/abs/2312.05230)

- GenRM: [Generative Verifiers: Reward Modeling as Next-Token Prediction](https://arxiv.org/abs/2408.15240)

- Perspective: [Towards Building Specialized Generalist AI with System 1 and System 2 Fusion](https://arxiv.org/abs/2407.08642)

### MCTS

- [Improve Mathematical Reasoning in Language Models by Automated Process Supervision](https://arxiv.org/abs/2406.06592)

- [LightZero: A Unified Benchmark for Monte Carlo Tree Search in General Sequential Decision Scenarios](https://arxiv.org/abs/2310.08348), code: [opendilab/LightZero](https://github.com/opendilab/LightZero)

- MuZero: [Mastering Atari, Go, Chess and Shogi by Planning with a Learned Model](https://arxiv.org/abs/1911.08265), open-source impl: [werner-duvaud/muzero-general](https://github.com/werner-duvaud/muzero-general)

### Minecraft Agents

- [Voyager: An Open-Ended Embodied Agent with Large Language Models](https://arxiv.org/abs/2305.16291)

- [JARVIS-1: Open-World Multi-task Agents with Memory-Augmented Multimodal Language Models](https://arxiv.org/abs/2311.05997)

- [Optimus-1: Hybrid Multimodal Memory Empowered Agents Excel in Long-Horizon Tasks](https://arxiv.org/abs/2408.03615)

- [STEVE Series: Step-by-Step Construction of Agent Systems in Minecraft](https://arxiv.org/abs/2406.11247)

#### Massive Sampling / Generate-and-Test

- [Scaling LLM Test-Time Compute Optimally can be More Effective than Scaling Model Parameters](https://arxiv.org/abs/2408.03314)

- [Large Language Monkeys: Scaling Inference Compute

with Repeated Sampling](https://arxiv.org/abs/2407.21787), [code](https://github.com/ScalingIntelligence/large_language_monkeys), [blog](https://scalingintelligence.stanford.edu/blogs/monkeys/)

- [AlphaCode 2 Technical Report](https://storage.googleapis.com/deepmind-media/AlphaCode2/AlphaCode2_Tech_Report.pdf)

### World Models

- GameNGen: [Diffusion Models Are Real-Time Game Engines](https://arxiv.org/abs/2408.14837), [project page](https://gamengen.github.io/)

- [A Path Towards Autonomous Machine Intelligence](https://openreview.net/pdf?id=BZ5a1r-kVsf)

- [GAIA-1: A Generative World Model for Autonomous Driving](https://arxiv.org/abs/2309.17080)

- Latent space world-models: [Dreamer](http://arxiv.org/pdf/1912.01603), [V2](https://arxiv.org/abs/2010.02193), [V3](https://arxiv.org/abs/2301.04104), [DayDreamer](https://arxiv.org/pdf/2206.14176)

- [World Models](https://arxiv.org/abs/1803.10122), web: [project page](https://worldmodels.github.io/)

- [Neural Assets: 3D-Aware Multi-Object Scene Synthesis with Image Diffusion Models](https://arxiv.org/abs/2406.09292)

### Neuro-Symbolic Approaches

- [HYSYNTH: Context-Free LLM Approximation for Guiding Program Synthesis](https://arxiv.org/abs/2405.15880)

- [SymbolicAI: A framework for logic-based approaches combining generative models and solvers](https://arxiv.org/abs/2402.00854), Library: [ExtensityAI/symbolicai](https://github.com/ExtensityAI/symbolicai)

- [DreamCoder: Growing generalizable, interpretable knowledge with wake-sleep Bayesian program learning](https://arxiv.org/abs/2006.08381)

- [A Neuro-vector-symbolic Architecture for Solving Raven's Progressive Matrices](https://arxiv.org/abs/2203.04571)

- Reasoning proofs generated by Prolog: [Neuro-Symbolic Integration Brings Causal and Reliable Reasoning Proofs](https://arxiv.org/abs/2311.09802), [Code](https://github.com/DAMO-NLP-SG/CaRing)

- VerityMath: [Advancing Mathematical Reasoning by Self-Verification Through Unit Consistency](https://arxiv.org/abs/2311.07172), [Code](https://github.com/vernontoh/VerityMath)

### Math

- AlphaGeometry: [Solving olympiad geometry without human demonstrations](https://www.nature.com/articles/s41586-023-06747-5)

- [Hologram Reasoning for Solving Algebra Problems with Geometry Diagrams](https://arxiv.org/abs/2408.10592)

- [Metacognitive Capabilities of LLMs: An Exploration in Mathematical Problem Solving](https://arxiv.org/abs/2405.12205)

### Active Inference

- [From pixels to planning: scale-free active inference](https://arxiv.org/abs/2407.20292)

## Prompting Techniques

- Surveys:

   - (Jul 2024) [The Prompt Report: A Systematic Survey of Prompting Techniques](https://arxiv.org/abs/2406.06608)

   - (Feb 2024) [A Systematic Survey of Prompt Engineering in Large Language Models: Techniques and Applications](https://arxiv.org/abs/2402.07927)

   - Prompt Engineering Guide [Prompting Techniques](https://www.promptingguide.ai/techniques)

   - [Prompting Fundamentals and How to Apply them Effectively](https://eugeneyan.com/writing/prompting/) by Eugene Yan

- Chain-of-Thoughts (COT): [Paper](https://arxiv.org/abs/2201.11903)

- Tree-of-Thoughts (ToT): [Paper](https://arxiv.org/pdf/2305.10601), impl: [Strategic Debate](https://github.com/zbambergerNLP/strategic-debate-tot)

- Graph-of-Thoughts (GoT): [Paper](https://arxiv.org/abs/2308.09687), [code](https://github.com/spcl/graph-of-thoughts)

- Algorithm of Thoughts (AoT): [Paper](https://arxiv.org/abs/2308.10379)

- Chain-of-Verification (CoVe/CoV): [Paper](https://arxiv.org/abs/2309.11495)

- Mixture-of-Agents (MoA): [Paper](https://arxiv.org/abs/2406.04692)

- Tool-Integrated Reasoning (ToRA / TIR): [Paper](https://arxiv.org/abs/2309.17452)

- Program of Thoughts (PoT): [Paper](https://arxiv.org/abs/2211.12588)

- Buffer of Thoughts (BoT): [Paper](https://arxiv.org/abs/2406.04271)

- Chain of Code (CoC): [Paper](https://arxiv.org/abs/2312.04474)

- Thought of Search (ToS): [Paper](https://arxiv.org/abs/2408.11326)

- Re-Reading the question as input (RE2): [Paper](https://arxiv.org/pdf/2309.06275)

## Mechanistic Interpretability

- [Geometric Space of Hierarchical Concepts in LLM](https://arxiv.org/abs/2406.01506)

## Blog Posts

- DeepMind [AlphaProof and AlphaGeometry 2](https://deepmind.google/discover/blog/ai-solves-imo-problems-at-silver-medal-level/)

- [Getting 50% (SoTA) on ARC-AGI with GPT-4o](https://www.lesswrong.com/posts/Rdwui3wHxCeKb7feK/getting-50-sota-on-arc-agi-with-gpt-4o), code: [rgreenblatt/arc_draw_more_samples_pub](https://github.com/rgreenblatt/arc_draw_more_samples_pub)

- Schmidhuber: [Artificial Curiosity & Creativity](https://people.idsia.ch/~juergen/artificial-curiosity-since-1990.html)

- synthesis.ai: [Do Androids Dream? World Models in Modern AI](https://synthesis.ai/2024/07/02/do-androids-dream-world-models-in-modern-ai/)

- [Our Transformers Code Agent beats the GAIA benchmark!](https://huggingface.co/blog/beating-gaia)

- Lil'Log [LLM Powered Autonomous Agents](https://lilianweng.github.io/posts/2023-06-23-agent/) (Jun 2023 )

- BAIR Blog: [The Shift from Models to Compound AI Systems](https://bair.berkeley.edu/blog/2024/02/18/compound-ai-systems/)

- Microsoft Research [Tracing the path to self-adapting AI agents](https://www.microsoft.com/en-us/research/blog/tracing-the-path-to-self-adapting-ai-agents/)

- [LLMs develop their own understanding of reality as their language abilities improve](https://news.mit.edu/2024/llms-develop-own-understanding-of-reality-as-language-abilities-improve-0814), [Emergent Representations Paper](https://arxiv.org/abs/2305.11169)

## Graph Neural Networks

- Distill [A Gentle Introduction to Graph Neural Networks](https://distill.pub/2021/gnn-intro/) (2021)

- [Geometric Deep Learning - Grids, Groups, Graphs, Geodesics, and Gauges](https://geometricdeeplearning.com/)

### Complex Logical Query Answering (CQLA)

Answering logical queries over Incomplete Knowledge Graphs. Aspirationally this requires combining sparse symbolic index collation (SQL, SPARQL, etc) and dense vector search, preferably in a differentiable manner.

- [Neural Graph Reasoning: Complex Logical Query Answering Meets Graph Databases](https://arxiv.org/abs/2303.14617)

- [Adapting Neural Link Predictors for Data-Efficient Complex Query Answering](https://openreview.net/forum?id=1G7CBp8o7L&referrer=%5Bthe%20profile%20of%20Erik%20Arakelyan%5D(%2Fprofile%3Fid%3D~Erik_Arakelyan1))

- [Generalizing Knowledge Graph Embedding with Universal Orthogonal Parameterization](https://openreview.net/forum?id=Sv4u9PtvT5)

- [Knowledge Sheaves: A Sheaf-Theoretic Framework for Knowledge Graph Embedding](https://arxiv.org/abs/2110.03789)

- [Wasserstein-Fisher-Rao Embedding: Logical Query Embeddings with Local Comparison and Global Transport](https://arxiv.org/abs/2305.04034)

- [GammaE: Gamma Embeddings for Logical Queries on Knowledge Graphs](https://arxiv.org/abs/2210.15578)

- [Soft Reasoning on Uncertain Knowledge Graphs](https://arxiv.org/abs/2403.01508)

### Inductive Reasoning over Heterogeneous Graphs

Similar to the regular CQLA, but with the emphasis on the "Inductive Setting" - i.e. querying over new, unseen during training nodes, edge types or even entire graphs. The latter part is interesting as it relies on the higher order "relations between relations" structure, connecting KG inference to Category Theory.

- [Zero-shot Logical Query Reasoning on any Knowledge Graph](https://arxiv.org/abs/2404.07198)

- [Extending Transductive Knowledge Graph Embedding Models for Inductive Logical Relational Inference](https://arxiv.org/abs/2309.03773)

- [Neural-Symbolic Models for Logical Queries on Knowledge Graphs](https://arxiv.org/abs/2205.10128)

- [InGram: Inductive Knowledge Graph Embedding via Relation Graphs](https://arxiv.org/abs/2305.19987)

### Neural Algorithmic Reasoning (NAR)

Initially attempted back in 2014 with general-purpose but unstable Neural Turing Machines, modern NAR approaches limit their scope to making GNN-based "Algorithmic Processor Networks" which learn to mimic classical algorithms on synthetic data and can be deployed on noisy real-world problems by sandwiching their frozen instances inside Encoder-Processor-Decoder architecture.

- [Neural Turing Machines, 2014](https://arxiv.org/abs/1410.5401)

- [A Generalist Neural Algorithmic Learner](https://openreview.net/forum?id=FebadKZf6Gd)

- [Transformers meet Neural Algorithmic Reasoners](https://arxiv.org/abs/2406.09308)

- [Recursive Algorithmic Reasoning](https://arxiv.org/abs/2307.00337)

- [Dual Algorithmic Reasoning](https://arxiv.org/abs/2302.04496)

- [Learning to Configure Computer Networks with Neural Algorithmic Reasoning](https://arxiv.org/abs/2211.01980)

## Grokking

- [Deep Networks Always Grok and Here is Why](https://imtiazhumayun.github.io/grokking/)

- [Grokfast: Accelerated Grokking by Amplifying Slow Gradients](https://arxiv.org/abs/2405.20233), [review post](https://x.com/_clashluke/status/1820810798693818761) by [Lucas Nestler](https://x.com/_clashluke)

- [Grokking: Generalization Beyond Overfitting on Small Algorithmic Datasets](https://arxiv.org/abs/2201.02177)

## Open-Source Agents & Agent Frameworks

- [gpt-researcher](https://github.com/assafelovic/gpt-researcher), [docs](https://docs.gptr.dev/docs/gpt-researcher/introduction)

- [open-interpreter](https://github.com/OpenInterpreter/open-interpreter), [docs](https://docs.openinterpreter.com/getting-started/introduction)

- [ADAS](https://github.com/ShengranHu/ADAS) (Automated Design of Agentic Systems)

- [AI-Scientist](https://github.com/SakanaAI/AI-Scientist.git)

- [Ollama_Agents](https://github.com/MikeyBeez/Ollama_Agents)

- [AgentK](https://github.com/mikekelly/AgentK)

- [Storm](https://github.com/stanford-oval/storm), [Paper](https://arxiv.org/abs/2402.14207)

- [crewAI](https://github.com/crewAIInc/crewAI), [docs](https://docs.crewai.com/)

- [AutoGPT](https://github.com/Significant-Gravitas/AutoGPT), [docs](https://docs.agpt.co/)

- [AutoGen](https://github.com/microsoft/autogen), [docs](https://microsoft.github.io/autogen/), [AutoGen Studio Paper](https://arxiv.org/abs/2408.15247)

- [Trace](https://github.com/microsoft/Trace), [docs](https://microsoft.github.io/Trace/intro.html), [Paper](https://arxiv.org/abs/2406.16218)

- [motleycrew](https://github.com/ShoggothAI/motleycrew), [docs](https://motleycrew.readthedocs.io/en/latest/)

- [langflow](https://github.com/langflow-ai/langflow), [docs](https://docs.langflow.org/)

## Algorithms

- [wake-sleep](https://en.wikipedia.org/wiki/Wake-sleep_algorithm)

- [novelty-search](https://algorithmafternoon.com/novelty/novelty_search_algorithm/)

### Weak Search Methods

Weak methods are general but don't use knowledge (heuristics) to guide the search process.

  - [depth-first-search](https://en.wikipedia.org/wiki/Depth-first_search) (DFS)

  - [breadth-first-search](https://en.wikipedia.org/wiki/Breadth-first_search) (BFS)

  - depth-limited-search, [iterative-deepening-depth-first-search](https://en.wikipedia.org/wiki/Iterative_deepening_depth-first_search) (IDDFS)

  - [generate-and-test](https://www.geeksforgeeks.org/generate-and-test-search/)

  - [hill-climbing](https://en.wikipedia.org/wiki/Hill_climbing) (borderline case between weak and strong methods)

### Strong Search Methods

- [best-first-search](https://en.wikipedia.org/wiki/Best-first_search)

- [A*](https://en.wikipedia.org/wiki/A*_search_algorithm)

- [beam-search](https://en.wikipedia.org/wiki/Beam_search)

- [monte-carlo-tree-search](https://en.wikipedia.org/wiki/Monte_Carlo_tree_search) (MCTS)

## Books

- [The Soar Cognitive Architecture](https://a.co/d/cXo8KFu), John E. Laird, MIT Press, 2019

- [How to Build a Brain: A Neural Architecture for Biological Cognition](https://academic.oup.com/book/6263) Chris Eliasmith, Oxford Series on Cognitive Models and Architectures, 2013

- [Active Inference: The Free Energy Principle in Mind, Brain, and Behavior](https://a.co/d/4OAX5dD), Thomas Parr, Giovanni Pezzulo, Karl J. Friston, MIT Press, 2022, [MLST Interview with Thomas Parr](https://www.youtube.com/watch?v=bk_xCikDUDQ)

- [Principles of Synthetic Intelligence PSI: An Architecture of Motivated Cognition](https://amzn.eu/d/46Xwq4s), Joscha Bach, Oxford Series on Cognitive Models and Architectures Book 4, 2009

- [Conscious Mind, Resonant Brain: How Each Brain Makes a Mind](https://a.co/d/5Hl3n7H), Stephen Grossberg, Oxford University Press, 2021

- [The Society of Mind](https://www.amazon.com/Society-Mind-Marvin-Minsky/dp/0671657135), Marvin Minsky, Simon & Schuster, 1986

- [Reinforcement Learning: An Introduction](http://incompleteideas.net/book/the-book-2nd.html) 2nd Edition, Sutton & Barto, MIT Press, 2018

- [Mathematical Foundations of Reinforcement Learning](https://github.com/MathFoundationRL/Book-Mathematical-Foundation-of-Reinforcement-Learning), Shiyu Zhao, open course on github + [video lectures](https://www.youtube.com/playlist?list=PLEhdbSEZZbDaFWPX4gehhwB9vJZJ1DNm8)

- [Natural Language Cognitive Architecture](https://github.com/daveshap/NaturalLanguageCognitiveArchitecture), David Shapiro, 2022, open source copy

- [An Introduction to Universal Artificial Intelligence](https://a.co/d/b61JiOh), Marcus Hutter, David Quarel, Elliot Catt, CRC Press, 2024 - AIXI, [Slides](http://hutter1.net/ai/suaibook.pdf), [Video](https://cartesiancafe.podbean.com/e/marcus-hutter-universal-artificial-intelligence-and-solomonoff-induction/)

## Biologically Inspired Approaches

Diverse approaches some of which tap into classical PDE systems of biological NNs, some concentrate on Distibuted Sparse Representations (by default non-differentiable), others draw inspiration from Hippocampal Grid Cells, Place Cells, etc. Biological systems surpass most ML methods for Continual and Online Learning, but are hard to implement efficienly on GPU.

- Ogma Sparse Predictive Hierarchies (SPH): [whitepaper](https://ogma.ai/sph-technology-description/)

- [The Tolman-Eichenbaum Machine: Unifying space and relational memory through generalisation in the hippocampal formation](https://www.biorxiv.org/content/10.1101/770495v2) (TEM), [TEM-t](https://arxiv.org/abs/2112.04035)

- [Arousal as a universal embedding for spatiotemporal brain dynamics](https://www.biorxiv.org/content/10.1101/2023.11.06.565918v2)

- [Sparse Distributed Memory is a Continual Learner](https://openreview.net/forum?id=JknGeelZJpHP)

- [Computation with Sequences of Assemblies in a Model of the Brain](https://proceedings.mlr.press/v237/dabagia24a.html)

### Dense Associative Memory

Dense Associative Memory is mainly represented by Modern Hopfield Networks (MHN), which [can be viewed](https://arxiv.org/abs/2008.02217) as a generalized Transformers capable of storing queries, keys and values explicitly (as in Vector Databases) and running recurrent retrival by energy minimization ([relating them](https://openreview.net/forum?id=B1BL9go65H&referrer=%5Bthe%20profile%20of%20Judy%20Hoffman%5D(%2Fprofile%3Fid%3D~Judy_Hoffman1)) to Diffusion models). Application for Continual Learning is possible when combined with uncertainty quantification and differentiable top-k selection.

- [xLSTM repository](https://github.com/NX-AI/xlstm)

- [CAMELoT: Towards Large Language Models with Training-Free Consolidated Associative Memory](https://arxiv.org/abs/2402.13449)

- [Energy Transformer](https://proceedings.neurips.cc/paper_files/paper/2023/hash/57a9b97477b67936298489e3c1417b0a-Abstract-Conference.html)

- [Memorization and consolidation in associative memory networks](https://openreview.net/forum?id=hXef89mdlH)

- [Simplicial Hopfield networks](https://openreview.net/forum?id=_QLsH8gatwx)

## Continual Learning

- [MagMax](https://arxiv.org/abs/2407.06322)

## Software Tools & Libraries

- [paul-gauthier/aider](https://github.com/paul-gauthier/aider)

- [claude-engineer](https://github.com/Doriandarko/claude-engineer)

- [continuedev/continue](https://github.com/continuedev/continue)

- [OpenHands](https://github.com/All-Hands-AI/OpenHands) (formerly OpenDevin)

- [princeton-nlp/SWE-agent](https://github.com/princeton-nlp/SWE-agent), [documentation](https://princeton-nlp.github.io/SWE-agent/)

- [meta-llama/llama-agentic-system](https://github.com/meta-llama/llama-agentic-system)

- [stanfordnlp/dspy]( https://github.com/stanfordnlp/dspy), DSPy awesome list: [ganarajpr/awesome-dspy](https://github.com/ganarajpr/awesome-dspy), [paper](https://arxiv.org/abs/2310.03714)

- [InternLM/lagent](https://github.com/InternLM/lagent) - lightweight framework for building LLM-based agents

## Commercial Offerings

- Software Engineering

  - [Devin](https://preview.devin.ai/)

  - [Cursor](https://www.cursor.com/)

  - [copilot-workspace](https://githubnext.com/projects/copilot-workspace)

  - [textgrad](https://textgrad.com/)

  - [Cosine Genie](https://cosine.sh/genie)

- Synthetic data generation

  - [glaive.ai](https://glaive.ai/) was used to generate CoT/Reflection data for [mattshumer/Reflection-Llama-3.1-70B](https://huggingface.co/mattshumer/Reflection-Llama-3.1-70B)

# Competitions & Benchmarks

- [AppWorld: A Controllable World of Apps and People for Benchmarking Interactive Coding Agents](https://arxiv.org/abs/2407.18901), web: [project page](https://appworld.dev/), gh: [stonybrooknlp/appworld](https://github.com/stonybrooknlp/appworld/), [Leaderboard](https://appworld.dev/leaderboard)

- [CRAB: Cross-environment Agent Benchmark for Multimodal Language Model Agents](https://arxiv.org/abs/2407.01511), gh: [camel-ai/crab](https://github.com/camel-ai/crab)

- [WebArena: A Realistic Web Environment for Building Autonomous Agents](https://arxiv.org/abs/2307.13854), web: [project page](https://webarena.dev/), [Leaderboard](https://docs.google.com/spreadsheets/d/1M801lEpBbKSNwP-vDBkC_pF7LdyGU1f_ufZb_NWNBZQ/edit?usp=sharing)

- [ARC-AGI](https://arcprize.org/arc): [Leaderboard](https://arcprize.org/leaderboard), [On the Measure of Intelligence](https://arxiv.org/abs/1911.01547)

- PlanBench: [Paper](https://arxiv.org/abs/2206.10498), gh: [karthikv792/LLMs-Planning](https://github.com/karthikv792/LLMs-Planning)

- [GAIA: a benchmark for General AI Assistants](https://arxiv.org/abs/2311.12983): [Leaderboard](https://gaia-benchmark-leaderboard.hf.space/)

- [StreamBench: Towards Benchmarking Continuous Improvement of Language Agents](https://arxiv.org/pdf/2406.08747), gh: [stream-bench/stream-bench](https://github.com/stream-bench/stream-bench)

- [VisualAgentBench: Towards Large Multimodal Models as Visual Foundation Agents](https://arxiv.org/abs/2408.06327)

- [ZebraLogic](https://huggingface.co/blog/yuchenlin/zebra-logic), [Leaderboard](https://huggingface.co/spaces/allenai/ZebraLogic)

### Code

- [SWE-bench](https://www.swebench.com/index.html), [SWE-bench Lite](https://www.swebench.com/lite.html)

- [BigCodeBench: The Next Generation of HumanEval](https://huggingface.co/blog/leaderboard-bigcodebench), [Leaderboard](https://huggingface.co/spaces/bigcode/bigcodebench-leaderboard)

- [SciCode: A Research Coding Benchmark Curated by Scientists](https://arxiv.org/abs/2407.13168), web: https://scicode-bench.github.io/

## Related Projects

- Nous Research [Open Reasoning Tasks](https://reasoning.nousresearch.com/), a list of reasoning tasks, gh: [NousResearch/Open-Reasoning-Tasks](https://github.com/NousResearch/Open-Reasoning-Tasks)

- [ARC-AGI Resources](https://docs.google.com/spreadsheets/d/1fR4cgjY1kNKN_dxiidBQbyT6Gv7_Ko7daKOjlYojwTY/edit?pli=1&gid=756763742#gid=756763742) Google table paper list by [ARC price](https://arcprize.org/)

## Youtube Content

- Channel: [David Shapiro](https://www.youtube.com/@DaveShap)

- [Artem Kirsanov: Engrams, Building Blocks of Memory in the Brain](https://www.youtube.com/watch?v=X5trRLX7PQY)

- Channel: [Edan Meyer on AI, ML & RL](https://www.youtube.com/@EdanMeyer), [Discrete vs. Continuous RL](https://www.youtube.com/watch?v=s8RqGlU5HEs) + [Paper](https://arxiv.org/abs/2312.01203)

- [MIT AGI: Cognitive Architecture (Nate Derbinsky)](https://www.youtube.com/watch?v=bfO4EkoGh40)

- Channel: [Thinking About Thinking](https://www.youtube.com/@ThoughtChannel/videos) (Mathematics of Neuroscience and AI)

- [Invariance and equivariance in brains and machines](https://youtu.be/xnhhp916JNU?si=VpiDQDbFafyNAlMW)

- [code_your_own_AI: The CORE IDEA of AI Agents Explained](https://youtu.be/xdAKa8jFx3g)

### Joscha Bach

- [Machine Consciousness](https://youtu.be/LlLbHm-bJQE)

- [Consciousness as a coherence-inducing operator](https://www.youtube.com/watch?v=qoHCQ1ozswA) Talk by Josha Bach at the [Models of Consciousness Conferences](https://models-of-consciousness.org/)

## Best LLM APIs

- [Anthropic Claude](https://docs.anthropic.com/en/api/getting-started)

- [together.ai](https://docs.together.ai/docs/introduction)

- [groq.com](https://console.groq.com)

- [openrouter.ai](https://openrouter.ai/)

## Novel model architectures

- [Mirasol3B: A Multimodal Autoregressive model for time-aligned and contextual modalities](https://arxiv.org/abs/2311.05698)

- [Memory3: Language Modeling with Explicit Memory](https://arxiv.org/abs/2407.01178)

- TTT: [Learning to (Learn at Test Time): RNNs with Expressive Hidden States](https://arxiv.org/abs/2407.04620), [Video](https://www.youtube.com/watch?v=I9Ghw2Z7Gqk)

- [TransformerFAM: Feedback attention is working memory](https://arxiv.org/pdf/2404.09173)

## Philosophy: Nature of Intelligence & Consciousness

- [A High Level Theory on the Nature of Intelligence and Consciousness](https://philarchive.org/rec/GARAHL)

## Biology / Neuroscience

- [The brain simulates actions and their consequences during REM sleep](https://www.biorxiv.org/content/10.1101/2024.08.13.607810v1)

- CSCG: [Clone-structured graph representations enable flexible learning and vicarious evaluation of cognitive maps](https://www.nature.com/articles/s41467-021-22559-5)

- [System-1 and System-2 realized within the Common Model of Cognition](https://www.researchgate.net/publication/365345252_System-1_and_System-2_realized_within_the_Common_Model_of_Cognition) (2022)

## Workshops

https://s2r-at-scale-workshop.github.io (NeurIPS 2024)

## How to contribute

 To share a link related to reasoning in AI systems that is missing here please create a pull request for this file. See [editing files](https://docs.github.com/en/repositories/working-with-files/managing-files/editing-files) in the github documentation.