The Scientist

Researchers Decode How Protein Language Models Think, Making AI More Transparent

By spreading out tightly packed information in neural networks, a new set of tools could make AI protein models easier to ...
GitHub

Train multi-step agents for real-world tasks using GRPO.

RULER (Relative Universal LLM-Elicited Rewards) eliminates the need for hand-crafted reward functions by using an LLM-as-judge to automatically score agent trajectories. Simply define your task in the ...

Alibaba’s Tongyi DeepResearch : Open-Source AI for Smarter Problem-Solving

Learn how Tongyi DeepResearch combines cutting-edge reasoning and open-source flexibility to transform advanced research ...

Tencent’s new AI technique teaches language models ‘parallel thinking’

The Parallel-R1 framework uses reinforcement learning to teach models how to explore multiple reasoning paths at once, ...
IEEE

Quantum-Train-Based Distributed Multi-Agent Reinforcement Learning

Abstract: In this paper, we introduce Quantum-Train-Based Distributed Multi-Agent Reinforcement Learning (Dist-QTRL), a novel approach to addressing the scalability challenges of traditional ...
IEEE

Low-rank Multi-modal Features Fusion Semantic Autoencoder For Zero-shot Learning

Abstract: In the context of the continuous increase of new things, zero-shot learning (ZSL) has been proposed to reduce recognition costs, with the goal of classifying or predicting new classes that ...

From stuck to scaled: How hyper-parallel AI training cuts iteration cycles 20X

When it comes to AI, many enterprises seem to be stuck in the prototype phase. Teams can be constrained by GPU capacity and ...