Research monitor · Page 2

Homebrewed SLMs: Autoregressive and Masked Diffusion Language Models Built at Home for Portuguese

Compute & efficiency · preprint

Building a language model of one's own no longer requires a data centre. This tutorial trains one from scratch, tokenizer included, on one consumer GPU,

RFW-A: Agentic Procedural Generation of 400K Articulated Assets over 500 Categories

Agent systems · preprint

High-quality articulated assets are essential for embodied perception, simulation, and manipulation, yet constructing them at scale remains difficult. Manual annotation and procedural modeling require substantial

Textual Response-Aware Guidance for Zero-Shot Adversarial Robustness of Vision-Language Models

Evaluation & safety · preprint

Vision-language models (VLMs) exhibit strong zero-shot generalization but remain vulnerable to adversarial attacks. Existing adversarial fine-tuning methods often use pre-trained model guidance to preserve image--text

Wizard: Manipulate Any World in Real Time via Agentic Control

Agent systems · preprint

Recent advances in video world models have enabled the generation of increasingly realistic and interactive virtual environments, opening new possibilities for player-customizable gameplay. Agentic frameworks

Hybrid AI Patterns: Thirteen Ways to Couple Language Models with Rules, and How to Choose Among Them

AI models · preprint

Language models handle text well, and rule engines give conclusions with proofs. Systems that combine the two are usually described by example, and the same

Video Generation Meets Agent Harness: Agentic Video Generation, Agentic World Modeling and Recursive Self-Improvement

Agent systems · preprint

Video generation is shifting from end-to-end synthesis toward agentic generation, where planning, tool use, and feedback enable coherent, controllable multi-shot videos. In parallel, video world

Prefix Executability: Evaluating Tool-Using Agents Beyond Final Success

Agent systems · preprint

Tool-use benchmarks often score agents by final task success, but terminal outcomes can obscure whether the intermediate tool-call trajectory was executable. We propose prefix executability

An updated Model Spec for more capable and beneficial AI

AI models · provider_announcement

OpenAI’s updated Model Spec outlines five core values to guide more capable AI toward sound judgment, human agency, and faithfulness to human intent.

Securing the Robotics Supply Chain

Robotics & embodiment · preprint

This paper sets out what a sustained federal research and investment agenda would need to include to close that gap--not within a single budget cycle,

A Conveyor Belt Mechanism for Repositionable Magnetic Coupling Between Continuum Robots

Robotics & embodiment · preprint

This work proposes a novel conveyor belt mechanism enabling continuous, active repositioning of functional elements along the body of a tendon-driven continuum robot. Up to

SelfThinker: Self-Evolving Reasoning via Reliability-Verified Intrinsic Visual Thoughts

Multimodal AI · preprint

Unified multimodal models can both generate and understand images within a single set of weights, offering an opportunity to use image synthesis as training supervision

How to Loop Your Transformer: Loop Scaling, Specialization, and Parallel Inference

Compute & efficiency · preprint

Looping in transformer models reuses weights by repeatedly applying the same transformer, or parts of it, making the model deeper without increasing its parameter count.

EnvOS: Auditing Computer-Use Agents When the World Moves On : Phenomenon-Based Task Construction, Certification, and Evidence-Based Evaluation

Agent systems · preprint

Computer-use agents built on large language models (LLMs) work in software that can change while they act: an appointment moves, a policy is revised, a

Agentic process evidence

Agent systems · preprint

The way software gets built is changing dramatically. AI agents now create pull requests, select dependencies, refactor codebases, interpret test failures, and execute infrastructure changes

Announcing Project ProtoStar: Making Publicly Funded Science AI-Ready

AI for science · primary_technical_publication

Don't Lose the Thread: Empowering Long-Horizon LLM Agents with Cognitive Resource Self-Allocation

Agent systems · preprint

Agents powered by large language models (LLMs) have demonstrated remarkable progress in solving complex reasoning tasks. However, LLM agents often falter on long-horizon tasks due

LOGIV: Graph-Verified Execution with Causal Local Repair for Long-Horizon Robot Tasks

Agent systems · preprint

Agentic robots must turn a high-level goal into actions while keeping the physical conditions needed for future steps valid. A completed action does not ensure

Gym-Anything for Robotics: Scaling physical simulations with digital agents

Agent systems · preprint

Realistic, fast robotics simulations could enable large-scale training of embodied AI models that transfer to the real world. However, despite decades of progress, available simulations

Agentic LLM reasoning in pediatric diabetes outcome prediction from multimodal medical records

Agent systems · preprint

Large language model (LLM) agents are proposed for clinical prediction but are untested in pediatrics, where cohorts are small and outcomes are uncommon. Using records

Retrieval as Traversal: Value-Indexed Memory for Stateful Long-Context Agents

Agent systems · preprint

Long-context models repeatedly solve closely related retrieval problems. Attention or global retrieval can retain large stores, but each forward pass again searches for what matters.

Agentic University: Collaborative Agents for Personalized Learning

Agent systems · preprint

Large language models support personalized tutoring, content generation, and adaptive learning paths. However, capability development requires these resources and interactions to form a coherent learning

DolphinRouting: Cluster-Level Binary Routing for Disaggregated LLM Inference Serving

Compute & efficiency · preprint

Prefill-decode (PD) disaggregation isolates the two phases of large language model serving, but request-level path selection must account for prefill work, cache reuse, and decode

The Athlete as a Scientist of One: Agentic Bayesian Experimental Design for Individualized Physiology from Consumer Wearables

Agent systems · preprint

Large language model (LLM) agents for personal health read wearable data and recommend, but do not learn about the individual in a way that can

Reinforcement Learning Under Runtime Confounding: Policy Learning and Off-Policy Evaluation

Evaluation & safety · preprint

Reinforcement learning (RL) is increasingly used to train robust, adaptable systems across healthcare, e-commerce, and robotics. In these settings, legal, ethical, or logistical concerns may