Is this you? As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.
Claim your profile
Articles by RLXF Framework
MIT and Recursion Release Open-Source Boltz-2 AI Model Predicting Protein Binding Affinity Near-FEP Accuracy 1000x Faster
Daily market recaps with key events, stock movements, and global influences 14 posts • GPT (4.1 mini) Updated Jun 8, 2025, 02:04 AM MIT and Recursion Pharmaceuticals have jointly released Boltz-2, an open-source artificial intelligence model that significantly advances drug discovery by predicting protein structure and binding affinity with near free-energy perturbation (FEP) accuracy at speeds up to 1,000 times faster than traditional simulations.
Recursion and MIT Release Open-Source Boltz-2 AI Model Predicting Protein Structure and Binding Affinity 1,000× Faster With Near-FEP Accuracy
13 posts • GPT (4.1 mini) Updated Jun 6, 2025, 03:51 PM Recursion Pharmaceuticals and the Massachusetts Institute of Technology (MIT) have jointly released Boltz-2, an open-source artificial intelligence model that advances drug discovery by predicting protein structure and binding affinity with near free-energy perturbation (FEP) accuracy.
Recursion and MIT Unveil Open-Source Boltz-2 AI Model with Near-FEP Accuracy and 1000x Speed for Binding Affinity Prediction
5 posts • GPT (4.1 mini) Published Jun 6, 2025, 02:14 PM Recursion Pharmaceuticals, in collaboration with the Massachusetts Institute of Technology (MIT) and a research team led by Regina Barzilay, has announced the release of Boltz-2, an open-source artificial intelligence model designed to predict protein structure and binding affinity. This model represents a notable advancement in drug discovery by accurately measuring how strongly molecules bind, a key factor in pharmaceutical development.
Tsinghua University and Ant Group Release Open-Source AReaL-boba² Async RL System With 2.77× Faster Training for Scalable LLM Reasoning
Daily market recaps with key events, stock movements, and global influences 9 posts • GPT (4.1 mini) Updated Jun 6, 2025, 08:26 AM Tsinghua University and Ant Group have introduced AReaL-boba², an asynchronous reinforcement learning (RL) system designed to enhance large language model (LLM) reasoning. The open-source framework offers scalability from single GPUs to large clusters and achieves approximately 2.77 to 2.8 times faster training speeds compared to synchronous RL baselines.
Ant Group and Tsinghua Launch Open-Source AReaL-boba² Async RL System With 2.77× Faster Training and Agentic RL Support
Daily market recaps with key events, stock movements, and global influences 5 posts • GPT (4.1 mini) Published Jun 5, 2025, 09:49 AM Researchers from Ant Group and Tsinghua University have developed AReaL-boba², an asynchronous reinforcement learning (RL) system designed to accelerate training for large language model (LLM) reasoning.
Transformer Models Achieve 96% Accuracy in Chemical Structure Prediction, Extend Protein Language Models to 61,000-Amino Acid Viral Genomes with Venus-MAXWELL and GeLLM4O-C | DeepNewz Bio
8 posts • GPT (4.1 mini) Published Jun 3, 2025, 06:07 PM Recent advances in machine learning and protein language models (PLMs) have led to notable developments in computational chemistry, molecular biology, and biomedical AI. A transformer-based multimodal multitask language model has achieved human-level accuracy in automated chemical structure elucidation from spectroscopic data, reaching up to 96% Top-1 accuracy.
CMU Develops INTUITOR and RENT Methods Using Internal Feedback for Self-Training LLMs on Math Reasoning Tasks
Daily market recaps with key events, stock movements, and global influences 10 posts • GPT (4.1 mini) Updated May 30, 2025, 11:08 PM Recent research from Carnegie Mellon University (CMU) and other institutions has introduced a new paradigm in training large language models (LLMs) called Reinforcement Learning from Internal Feedback (RLIF).
New Reinforcement Learning Method Enables Qwen 2.5 Model to Self-Improve With 60% Code Gain, Matches RLHF on Math, Scales to 14B Parameters | DeepNewz AI Modeling
Daily market recaps with key events, stock movements, and global influences 15 posts • GPT (4.1 mini) Updated May 29, 2025, 09:58 PM A new research paper titled "Learning to Reason without External Rewards," authored by teams from UC Berkeley and Yale University, proposes a novel reinforcement learning approach called Reinforcement Learning from Internal Feedback (RLIF).
Intuitor Project Advances LLM Reasoning With 60% Code Improvement; CellReasoner and Qwen Models Highlight New RL Approaches
Daily market recaps with key events, stock movements, and global influences 16 posts • GPT (4.1 mini) Published May 27, 2025, 08:23 PM Recent research in large language models (LLMs) and reinforcement learning (RL) has introduced novel approaches to enhance reasoning capabilities without relying on traditional external reward signals.
Andrew Ng, DeepLearningAI Launch Free Reinforcement Fine-Tuning Course Using GRPO Algorithm and System Prompt Learning
Daily market recaps with key events, stock movements, and global influences 6 posts • GPT (4.1 mini) Published May 21, 2025, 06:54 PM A new paradigm in large language model (LLM) development is emerging with the introduction of Reinforcement Fine-Tuning (RFT) powered by the Group Relative Policy Optimization (GRPO) algorithm.
eSPred and Xpressor Models Enhance Single-Cell RNA Sequencing with Cross-Attention and Multi-Scale Fine-Tuning
Daily market recaps with key events, stock movements, and global influences 4 posts • GPT (4.1 mini) Published May 20, 2025, 09:28 PM Recent developments in computational biology have introduced two advanced foundation models aimed at enhancing biological data analysis across multiple scales. The first model, eSPred, is a customized foundation model specifically designed for single-cell RNA sequencing (scRNA-seq) predictive analysis.
Duke’s DPAC AI Uses Contrastive Learning and Simulated Annealing to Predict Protein-DNA Interactions Without Structural Data
4 posts • GPT (4.1 mini) Published May 20, 2025, 09:29 PM Researchers at Duke University have developed DPAC (DNA-Protein binding Alignment via Contrastive learning), a novel AI-driven framework for predicting and designing protein-DNA interactions using only sequence data. Unlike traditional methods that rely on complex three-dimensional structural information, DPAC employs a dual-tower model combining a Protein Language Model and a DNA Language Model.
Show More
loading
Actions
Is this you?
As a journalist, you can create a free Muck Rack account to customize your profile, list your contact preferences, and upload a portfolio of your best work.Get in touch with RLXF
Contact RLXF, search articles and posts on X, monitor coverage, and track replies from one place.
Learn more about Muck Rack