Find open-source science resources

Transform arXiv papers into Beamer slides using LLMs

Structure-aware protein language model using 3D structural vocabulary (Foldseek) for joint sequence-structure pretraining, achieving SOTA on protein engineering and fitness prediction benchmarks (ICML 2024, Westlake University & Repl)

Bilingual protein language model translating between protein sequence and structure, finetuned from ProtT5-XL on 17M AlphaFoldDB structures using Foldseek's 3Di structural alphabet, enabling sequence-to-structure prediction, structure-to-sequence inverse folding, and unified protein representation learning (RostLab, 310+ stars)

First benchmark for automatic video generation from scientific papers (NeurIPS 2025)

RFantibody is a pipeline for structure-based de novo antibody and nanobody design, integrating backbone design with RFdiffusion, sequence design with ProteinMPNN, and structure prediction with RoseTTAFold2. It provides a comprehensive toolset for generating and filtering high-quality antibody designs.

Interactive personal genome analysis toolkit using Claude Code and Python. Parses raw genotyping data from consumer DNA services and analyzes SNPs across 17 categories including health risks, pharmacogenomics, ancestry, and nutrition, with a terminal-style HTML dashboard.

Open-source platform for building, extending, and experimenting with scientific agents, providing modular agent construction tools and standardized evaluation pipelines for accelerating autonomous scientific discovery research (748+ stars, MIT License)

First benchmark evaluating LLMs' ability to rediscover scientific laws through interactive experimentation across 324 tasks in 12 physics domains, featuring memorization-resistant metaphysical shifts of canonical laws (HKUST)

Flow-matching protein folding model using only general-purpose transformer layers, scaled to 3B parameters and trained on 8.6M+ distilled structures; challenges the reliance on complex domain-specific architectures and supports PyTorch and MLX backends with model sizes from 100M to 3B parameters (985+ stars, MIT License)

Physics-informed neural networks

Azure Semantic Kernel multi-agent PPT generation reference

Multi-agent system automatically transforming research papers into interactive AI agents with MCP server generation, tutorial auto-detection, and benchmark extraction (2.2K+ stars, MIT License, 2025)

Family of large language models for materials research via continued pretraining of LLaMA-2/3 on ~30B materials science tokens, outperforming commercial LLMs on materials science tasks while identifying "adaptation rigidity" in overtrained models; includes MatNLP benchmark and CIF crystal generation capabilities (IIT Delhi M3RG, MIT License)

Evaluating multimodal autonomous agents in realistic scientific workflows across real scientific software environments (KAlgebra, Celestia, Grass GIS, Lean 4, etc.) with VM-based evaluation infrastructure and agent trajectories

Open-source toolkit and benchmark for learning-based theorem proving in Lean, providing programmatic Lean interaction, a 98K+ theorem dataset extracted from 217 Lean projects, and ReProver—the first retrieval-augmented LLM-based theorem prover for Lean—with reproducible training pipelines underpinning much subsequent Lean prover research (Caltech & NVIDIA, NeurIPS 2023 Outstanding Paper, Datasets & Benchmarks)

Discrete diffusion framework for generative protein sequence design over evolutionary-scale databases, supporting unconditional generation, evolutionary-guided conditional design, motif scaffolding, and intrinsically disordered region generation through order-agnostic autoregressive diffusion, enabling sequence-only protein design without structural priors (Microsoft Research, Nature Communications 2024)

AI-human collaborative research platform where a human researcher works with a team of LLM agents via team and individual meetings to perform scientific research; demonstrated by designing new SARS-CoV-2 nanobodies with wet-lab validation

Deep equivariant generative model predicting ligand-specific protein-ligand complex structures with dynamic receptor conformational flexibility, enabling accurate docking for flexible protein targets

Cross-modal self-supervised foundation model for galaxies by Polymathic AI, jointly embedding multi-band galaxy imaging and optical spectra into a shared latent space to enable zero/few-shot redshift estimation, galaxy property prediction, morphology classification, and cross-modal similarity search (MNRAS Letters 2024)

Generative AI framework for inverse design of 3D RNA structure and function using geometric deep learning, learning design rules from 3D structures to capture complex tertiary interactions (pseudoknots, non-canonical base pairs) with expert-level accuracy for designing functional RNAs including aptamers and ribozymes (bioRxiv 2025)

Discovering interpretable features in protein language models via sparse autoencoders, enabling mechanistic understanding of PLM representations for protein engineering and design (288+ stars, MIT License)

LLM agent system synthesizing Wikipedia-like long-form research articles from scratch through multi-perspective question asking, web retrieval, and citation-grounded report generation, with Co-STORM extension for collaborative human-LLM knowledge curation conversations (Stanford OVAL, NAACL 2024 & EMNLP 2024)

Scientific equation discovery and symbolic regression using LLMs, combining code generation with evolutionary search (ICLR 2025 Oral)

AI agent for therapeutic reasoning across a universe of tools, achieving 92.1% accuracy in drug reasoning and outperforming GPT-4o by 25.8% (Harvard MIMS, 2025)

Multi-modal geospatial ML platform for agriculture and sustainability, fusing satellite imagery (RGB, SAR, multispectral), drone imagery, weather data, and sensor data for crop identification, carbon footprint estimation, and microclimate prediction (Microsoft Research, MIT License)

Therapeutics Data Commons: 66 AI-ready datasets across 22 drug discovery tasks with 29 leaderboards, covering target identification, molecular generation, ADMET prediction, and clinical trial outcomes (Harvard MIMS, NeurIPS 2021/2024)

AI agent for biological discovery and research automation

Curated list of large weather models for AI Earth science

Curated scientific LLM papers (260+ models)

Docling-powered parsing with UI/CLI demonstration for rapid prototyping

Universal 3D molecular pretraining framework with 209M conformations, scaling to 1.1B parameters (Uni-Mol2) on 800M conformations for molecular property prediction, docking, and quantum chemistry (ICLR 2023, NeurIPS 2024)

RNA foundation model trained on millions of RNA sequences for generalist RNA sequence understanding, enabling downstream structure prediction, function annotation, and representation learning for non-coding RNAs (ml4bio, 372+ stars)

State-of-the-art pretrained language models for proteins trained on thousands of GPUs and Google TPUs using Transformer architectures, enabling protein property prediction, feature extraction, and transfer learning across diverse downstream tasks (1.3K+ stars, MIT, 2020-2026)

Diffusion-based molecular docking achieving SOTA blind docking performance, treating ligand pose prediction as generative diffusion over SE(3), with DiffDock-L update for improved generalization (MIT CSAIL, ICLR 2023)

Unified benchmarking framework for protein representation learning, providing standardized interfaces for pre-training and diverse downstream tasks including structure prediction, fitness prediction, and property prediction across multiple protein datasets and model architectures (ICLR 2024, 273+ stars, MIT License)

PyTorch implementation of neural ODEs

Neural optical understanding for academic documents, transforms scientific PDFs to Markdown with mathematical formula support

Microsoft's AI-powered ab initio biomolecular dynamics simulation achieving quantum-mechanical accuracy for proteins with 10,000+ atoms, orders of magnitude faster than DFT using protein fragmentation and ML force fields (Nature 2024)

Versatile multi-temporal geospatial foundation model for Earth observation, built on a ViT-based masked autoencoder with 3D spatiotemporal patch embeddings and geolocation/temporal metadata encoding; pretrained on 4.2M global time-series samples from NASA's Harmonized Landsat and Sentinel-2 archive at 30m resolution, with 300M/600M parameter variants and fine-tuning configs for flood detection, wildfire scar, landslide detection, crop segmentation, land cover, and biomass estimation (258+ stars, MIT License)

Equivariant graph attention Transformer (ICLR2023)

Extension of ProteinMPNN for protein sequence design in the context of small-molecule ligands, metal ions, and nucleic acids, enabling binding site engineering and co-factor redesign (Baker Lab)

Geometric deep learning model predicting transcriptional outcomes of novel single- and multi-gene perturbations using gene–gene knowledge graphs, 40% higher precision than prior methods on combinatorial perturbation prediction (Stanford, Nature Biotechnology 2024)

LLM for scientific research papers

Kolmogorov-Arnold Networks with learnable activation functions on edges instead of fixed node activations, achieving strong performance in function fitting, PDE solving, and scientific discovery with enhanced interpretability as an alternative to MLPs (MIT, 16.3K+ stars, 2024)

LLM agents across scientific domains

Large-scale biomolecular instruction dataset for chemistry/biology LLMs (ICLR2024)

Chemical language model

Materials informatics benchmark

Web application for LLM-assisted manuscript review and annotation

Deep learning-based protein sequence design (inverse folding) from backbone structures, achieving 52.4% sequence recovery vs 32.9% for Rosetta, core tool in modern protein design pipelines (Baker Lab, Science 2022)