BeamPERL: PE-RLVR-FT for Beam Mechanics Problem-Solving
AI & ML interests
PI: Markus J. Buehler, MIT. Our research focus on developing a new paradigm that designs materials from the molecular scale, using MD, ML and other methods.
Recent Activity
Papers
GraphAgents: Knowledge Graph-Guided Agentic AI for Cross-Domain Materials Design
Selective Imperfection as a Generative Framework for Analysis, Creativity and Discovery
Graph-native flow models (instruct to fine-tuned reasoning models)
PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking
-
PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking
Paper • 2410.12375 • Published • 5 -
In-situ graph reasoning and knowledge expansion using Graph-PReFLexOR
Paper • 2501.08120 • Published • 6 -
lamm-mit/PRefLexOR_ORPO_DPO_EXO_10242024
Text Generation • 4B • Updated • 1 -
lamm-mit/PRefLexOR_ORPO_DPO_EXO_REFLECT_10222024
Text Generation • 4B • Updated • 8 • 3
Cephalo is a series of multimodal vision large language models (V-LLMs) designed to integrate visual and linguistic reasoning in materials science.
Collection of fine-tuned Stable Diffusion models (SD2.x, SD-XL, SD3) to incorporate biological design cues, here exemplified for leaf microstructures.
Bioinspired123D is a generative agentic framework that translates natural language descriptions of biological structure into executable Blender code.
Collection of diffusion LLMs adapted for materials science and biology.
We present an approach to modifying Transformer architectures by integrating graph-aware relational reasoning into the attention mechanism.
-
lamm-mit/Llama-3.2-3B-Instruct-Sparse-GIN-orca-math-word-problems
Updated • 1 -
lamm-mit/Llama-3.2-3B-Instruct-Sparse-GIN-logic
2B • Updated • 1 • 1 -
lamm-mit/Llama-3.2-3B-Instruct-Sparse-GIN-bio
Updated • 1 • 1 -
Graph-Aware Isomorphic Attention for Adaptive Dynamics in Transformers
Paper • 2501.02393 • Published • 7
-
Cephalo: Multi-Modal Vision-Language Models for Bio-Inspired Materials Analysis and Design
Paper • 2405.19076 • Published • 2 -
X-LoRA: Mixture of Low-Rank Adapter Experts, a Flexible Framework for Large Language Models with Applications in Protein Mechanics and Design
Paper • 2402.07148 • Published • 6 -
ProtAgents: Protein discovery via large language model multi-agent collaborations combining physics and machine learning
Paper • 2402.04268 • Published -
MechAgents: Large language model multi-agent collaborations can solve mechanics problems, generate new data, and integrate knowledge
Paper • 2311.08166 • Published • 2
A set of LLMs fine-tuned on biological materials, mechanics, and materials science applications.
BeamPERL: PE-RLVR-FT for Beam Mechanics Problem-Solving
Bioinspired123D is a generative agentic framework that translates natural language descriptions of biological structure into executable Blender code.
Graph-native flow models (instruct to fine-tuned reasoning models)
Collection of diffusion LLMs adapted for materials science and biology.
We present an approach to modifying Transformer architectures by integrating graph-aware relational reasoning into the attention mechanism.
-
lamm-mit/Llama-3.2-3B-Instruct-Sparse-GIN-orca-math-word-problems
Updated • 1 -
lamm-mit/Llama-3.2-3B-Instruct-Sparse-GIN-logic
2B • Updated • 1 • 1 -
lamm-mit/Llama-3.2-3B-Instruct-Sparse-GIN-bio
Updated • 1 • 1 -
Graph-Aware Isomorphic Attention for Adaptive Dynamics in Transformers
Paper • 2501.02393 • Published • 7
PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking
-
PRefLexOR: Preference-based Recursive Language Modeling for Exploratory Optimization of Reasoning and Agentic Thinking
Paper • 2410.12375 • Published • 5 -
In-situ graph reasoning and knowledge expansion using Graph-PReFLexOR
Paper • 2501.08120 • Published • 6 -
lamm-mit/PRefLexOR_ORPO_DPO_EXO_10242024
Text Generation • 4B • Updated • 1 -
lamm-mit/PRefLexOR_ORPO_DPO_EXO_REFLECT_10222024
Text Generation • 4B • Updated • 8 • 3
-
Cephalo: Multi-Modal Vision-Language Models for Bio-Inspired Materials Analysis and Design
Paper • 2405.19076 • Published • 2 -
X-LoRA: Mixture of Low-Rank Adapter Experts, a Flexible Framework for Large Language Models with Applications in Protein Mechanics and Design
Paper • 2402.07148 • Published • 6 -
ProtAgents: Protein discovery via large language model multi-agent collaborations combining physics and machine learning
Paper • 2402.04268 • Published -
MechAgents: Large language model multi-agent collaborations can solve mechanics problems, generate new data, and integrate knowledge
Paper • 2311.08166 • Published • 2
Cephalo is a series of multimodal vision large language models (V-LLMs) designed to integrate visual and linguistic reasoning in materials science.
Collection of fine-tuned Stable Diffusion models (SD2.x, SD-XL, SD3) to incorporate biological design cues, here exemplified for leaf microstructures.
A set of LLMs fine-tuned on biological materials, mechanics, and materials science applications.