SIGNAL
Tracking the global AI frontier — labs · research · agents · policy
Frontier Signal
⚖

Policy & World

AI regulation and governance across the EU, US, China, UK and Asia: the AI Act, national strategies, safety institutes and global policy moves.

AI regulation and governance across the EU, US, China, UK and Asia: the AI Act, national strategies, safety institutes and global policy moves.

Latest in Policy & World

30 stories
Not All Claims Are Equally Risky: FACTOR for Adaptive Verification in Factual Long-Form Generation
Research

Not All Claims Are Equally Risky: FACTOR for Adaptive Verification in Factual Long-Form Generation

Large Language Models (LLMs) generate fluent long-form text, however, often add unsupported factual claims. Existing verification techniques improve factuality …
PolicyAlign: Direct Policy-Based Safety Alignment for Large Language Models
Research

PolicyAlign: Direct Policy-Based Safety Alignment for Large Language Models

Safety alignment of large language models (LLMs) typically depends on high-quality supervision data, such as safe demonstrations or preference pairs. However, i…
Selective Capability Unlearning in End-to-End Spoken Language Understanding
Research

Selective Capability Unlearning in End-to-End Spoken Language Understanding

Modern spoken language understanding (SLU) systems are increasingly deployed in real-world settings, where specific functionalities may need to be removed due t…
Select-to-Act: Hierarchical Reinforcement Learning via Adaptive Language Guidance
Research

Select-to-Act: Hierarchical Reinforcement Learning via Adaptive Language Guidance

Reinforcement Learning (RL) has been widely applied to sequential decision-making, yet it often suffers from poor sample efficiency due to costly interactions w…
Low Variance Trust Region Optimization with Independent Actors and Sequential Updates in Cooperative Multi-agent Reinforcement Learning
Research

Low Variance Trust Region Optimization with Independent Actors and Sequential Updates in Cooperative Multi-agent Reinforcement Learning

Cooperative multi-agent reinforcement learning assumes each agent shares the same reward function and can be trained effectively using the Trust Region framewor…
RecallRisk-BERT: A Multi-Task Framework for Post-Report Medical Device Recall Triage
Research

RecallRisk-BERT: A Multi-Task Framework for Post-Report Medical Device Recall Triage

Medical device recalls are a critical regulatory mechanism for protecting patient safety. The growing volume of FDA recall records presents challenges in post-r…
Production-grade AI agents for financial compliance: Lessons from Stripe
Practice

Production-grade AI agents for financial compliance: Lessons from Stripe

In this post, you learn how Stripe built a production-grade AI agent system for financial compliance. We cover the technical architecture of Stripe’s ReAct agen…
Turning sustainability compliance into a competitive edge
Practice

Turning sustainability compliance into a competitive edge

How financial institutions can turn mandatory reporting of financed emissions into a growth opportunity.…
LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation-assisted interpretation
Research

LCAi: Life Cycle Assessment with big data fusion and retrieval-augmented generation-assisted interpretation

The interpretation phase of life cycle assessment often lacks structured mechanisms for translating quantified improvement opportunities addressing environmenta…
Generative Retrieval via Diffusion Transformer with Metric-Ordered Sequence Training and Hybrid-Policy Preference Optimization
Research

Generative Retrieval via Diffusion Transformer with Metric-Ordered Sequence Training and Hybrid-Policy Preference Optimization

Embedding-based retrieval ranks items by their similarity to a query in a shared vector space and usually aims to return the highest-scoring items. In many prod…
A Process Harness for Uplifting Legacy Workflows to Agentic BPM: Design and Realization in CUGA FLO
Research

A Process Harness for Uplifting Legacy Workflows to Agentic BPM: Design and Realization in CUGA FLO

We introduce the process harness, a new mechanism for uplifting legacy workflows into Agentic Business Process Management (Agentic BPM) without replacing the un…
Designing Reward Signals for Portable Query Generation: A Case Study in Industrial Semantic Job Search
Research

Designing Reward Signals for Portable Query Generation: A Case Study in Industrial Semantic Job Search

Job-search platforms rely on low-bandwidth query interfaces that often fail to capture the high-dimensional complexity of candidate profiles. We present an end-…
Physics Question Scene Graph: Fine-grained Evaluation of Physical Plausibility in Text-to-Video Generation
Research

Physics Question Scene Graph: Fine-grained Evaluation of Physical Plausibility in Text-to-Video Generation

Video generation models are increasingly capable of producing realistic videos, but they still struggle to generate videos that follow basic physical laws. Comp…
Beyond Next-Observation Prediction: Agent-Authored World Modeling for Sequential Decision Making
Research

Beyond Next-Observation Prediction: Agent-Authored World Modeling for Sequential Decision Making

Recent studies on world modeling for Large Language Model (LLM) agents typically formulate the learning objective as next-observation prediction. However, this …
Low-Complexity Policy Tessellations in Structured Markov Decision Processes
Research

Low-Complexity Policy Tessellations in Structured Markov Decision Processes

We study optimal-policy geometry in structured Markov decision processes. While approximate dynamic programming and reinforcement learning typically approximate…
Constraint Tax in Open-Weight LLMs: An Empirical Study of Tool Calling Suppression Under Structured Output Constraints
Research

Constraint Tax in Open-Weight LLMs: An Empirical Study of Tool Calling Suppression Under Structured Output Constraints

Tool Calling and Structured Output are two core capabilities of modern Agent systems, yet their interaction under joint deployment conditions remains insufficie…
Semantic Consistency Policy Optimization for Reinforcement Learning of LLM Agents
Research

Semantic Consistency Policy Optimization for Reinforcement Learning of LLM Agents

Group-based reinforcement learning effectively post-trains LLM agents for long-horizon, sparse-reward tasks by deriving step-level credit from trajectory outcom…
Multi-Agent Goal Recognition with Team- and Goal-Conditioned Reinforcement Learning and Factorized Branch-and-Bound
Research

Multi-Agent Goal Recognition with Team- and Goal-Conditioned Reinforcement Learning and Factorized Branch-and-Bound

Multi-agent goal recognition asks an observer to jointly infer which agents act together and what each team is trying to achieve, so the hypothesis space grows …
An Introduction to Causal Reinforcement Learning
Research

An Introduction to Causal Reinforcement Learning

Causal inference provides a set of principles and tools that allow one to combine data and knowledge about an environment to reason with questions of counterfac…
Structural Kolmogorov-Arnold Convolutions: Learnable Function on the Values or the Filter Shape as Parameter-Efficient Alternative to Per-Edge Convolutional KANs
Research

Structural Kolmogorov-Arnold Convolutions: Learnable Function on the Values or the Filter Shape as Parameter-Efficient Alternative to Per-Edge Convolutional KANs

Convolutional Kolmogorov--Arnold Networks (KANs) replace the fixed weights of a convolutional kernel with learnable univariate functions. The dominant formulati…
Infinitesimal Causality
Research

Infinitesimal Causality

This paper introduces a categorical account of infinitesimal causality in Frobenius Markov categories equipped with tangent-bundle semantics. IDC captures the i…
Revealing Training Data Exposure in Vision Language Large Models via Parameter Gradients
Research

Revealing Training Data Exposure in Vision Language Large Models via Parameter Gradients

Vision-Language Large Models (VLLMs) trained on massive crawled corpora raise pressing copyright and data-provenance concerns. These concerns are particularly a…
Policy-as-Data: Learning Generalizable HOI Diffusion Models from Simulated Physics
Research

Policy-as-Data: Learning Generalizable HOI Diffusion Models from Simulated Physics

Synthesizing realistic Human-Object Interactions (HOI) is critical for creating embodied avatars and functional virtual environments. However, current data-driv…
SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning
Research

SingGuard: A Policy-Adaptive Multimodal LLM Guardrail with Dynamic Reasoning

Vision-language models (VLMs) are increasingly deployed in consumer, medical, financial, and enterprise applications. This broad deployment expands the safety s…
Have You Ever Seen Them? Entity-level Membership Inference through Interrogating Large Language Models
Research

Have You Ever Seen Them? Entity-level Membership Inference through Interrogating Large Language Models

Large Language Models (LLMs) raise growing concerns about privacy leakage and copyright compliance. Membership inference is a key tool for assessing such risks,…
Self-Evolution for Multi-Turn Tool-Calling Agents via Divergence-Point Preference Learning
Research

Self-Evolution for Multi-Turn Tool-Calling Agents via Divergence-Point Preference Learning

Multi-turn tool-using agents must coordinate long-horizon tool sequences while tracking dialogue state and policy constraints. Existing approaches often separat…
Build your own claw and agent harness with Microsoft Agent Framework
Agents

Build your own claw and agent harness with Microsoft Agent Framework

What does it take to build your own “claw” – a capable, CLI-style agent that can plan, use tools, remember things, and safely act on your beha…
Capturing growth in a fragmented world
Practice

Capturing growth in a fragmented world

CEOs who act nimbly and know where to place their bets can gain a competitive edge amid a shifting geopolitical landscape.…
Localizing Credit at the Divergence: Path-Conditioned Self-Distillation for LLM Reasoning
Research

Localizing Credit at the Divergence: Path-Conditioned Self-Distillation for LLM Reasoning

Reinforcement learning from verifiable rewards assigns a single scalar to each rollout, leaving token-level credit assignment underspecified in long reasoning t…
Auditing Machine Unlearning: A Systematic Research on Whether Models Truly Forget
Research

Auditing Machine Unlearning: A Systematic Research on Whether Models Truly Forget

Machine unlearning has been extensively studied in response to growing privacy concerns and regulatory requirements. However, auditing whether unlearning algori…