Gilead has partnered with digital health company Manta Cares to get tools for navigating lung and breast cancer care pathways to patients.
Co-creating a program theory and evaluability assessment for an Irish single-session, synchronous chat-based youth mental health intervention: implications for outcome evaluation
IntroductionSingle-session online synchronous chat offers immediate, anonymous, single-session support for young people. However, the drop-in format attracts a diverse population with urgent and varied needs, creating challenges for evaluation. Standardized outcome measures may not capture short-term changes, and randomized controlled trials may be ethically inappropriate. These constraints point to the value of theory-based evaluation approaches […]
Merck partners with Flagship startup Quotient to search for genetic clues to IBD
Quotient Therapeutics, a startup studying how tiny genetic typos cause some cells to get sick but allow others to remain healthy, has struck a partnership with Merck. The partnership will identify new drug targets for …
Gilead buys autoimmune disease biotech Ouro for $2.2bn
Gilead has agreed to buy Ouro and its T-cell engagers for autoimmune diseases in a $2.2bn deal that may also revive its alliance with Galapagos.
Immutrin secures $86M for antibody program
A British biotech based in Cambridge, England, raised $86 million to advance an antibody drug designed to deplete protein deposits in people with amyloidosis. The company, called Immutrin, was co-founded by Gregory Winter. He won …
Efficient Failure Management for Multi-Agent Systems with Reasoning Trace Representation
arXiv:2603.21522v1 Announce Type: cross Abstract: Large Language Models (LLM)-based Multi-Agent Systems (MASs) have emerged as a new paradigm in software system design, increasingly demonstrating strong reasoning and collaboration capabilities. As these systems become more complex and autonomous, effective failure management is essential to ensure reliability and availability. However, existing approaches often rely on per-trace reasoning, […]
Efficient Zero-Shot AI-Generated Image Detection
arXiv:2603.21619v1 Announce Type: cross Abstract: The rapid progress of text-to-image models has made AI-generated images increasingly realistic, posing significant challenges for accurate detection of generated content. While training-based detectors often suffer from limited generalization to unseen images, training-free approaches offer better robustness, yet struggle to capture subtle discrepancies between real and synthetic images. In this […]
ProMAS: Proactive Error Forecasting for Multi-Agent Systems Using Markov Transition Dynamics
arXiv:2603.20260v1 Announce Type: new Abstract: The integration of Large Language Models into Multi-Agent Systems (MAS) has enabled the so-lution of complex, long-horizon tasks through collaborative reasoning. However, this collec-tive intelligence is inherently fragile, as a single logical fallacy can rapidly propagate and lead to system-wide failure. Most current research re-lies on post-hoc failure analysis, thereby […]
Bounded Coupled AI Learning Dynamics in Tri-Hierarchical Drone Swarms
arXiv:2603.20333v1 Announce Type: cross Abstract: Modern autonomous multi-agent systems combine heterogeneous learning mechanisms operating at different timescales. An open question remains: can one formally guarantee that coupled dynamics of such mechanisms stay within the admissible operational regime? This paper studies a tri-hierarchical swarm learning system where three mechanisms act simultaneously: (1) local Hebbian online learning […]
On the Fragility of AI Agent Collusion
arXiv:2603.20281v1 Announce Type: cross Abstract: Recent work shows that pricing with symmetric LLM agents leads to algorithmic collusion. We show that collusion is fragile under the heterogeneity typical of real deployments. In a stylized repeated-pricing model, heterogeneity in patience or data access reduces the set of collusive equilibria. Experiments with open-source LLM agents (totaling over […]
Nemotron-Cascade 2: Post-Training LLMs with Cascade RL and Multi-Domain On-Policy Distillation
arXiv:2603.19220v2 Announce Type: replace-cross Abstract: We introduce Nemotron-Cascade 2, an open 30B MoE model with 3B activated parameters that delivers best-in-class reasoning and strong agentic capabilities. Despite its compact size, its mathematical and coding reasoning performance approaches that of frontier open models. It is the second open-weight LLM, after DeepSeekV3.2-Speciale-671B-A37B, to achieve Gold Medal-level performance […]
LLMServingSim 2.0: A Unified Simulator for Heterogeneous and Disaggregated LLM Serving Infrastructure
arXiv:2602.23036v2 Announce Type: replace-cross Abstract: Large language model (LLM) serving infrastructures are undergoing a shift toward heterogeneity and disaggregation. Modern deployments increasingly integrate diverse accelerators and near-memory processing technologies, introducing significant hardware heterogeneity, while system software increasingly separates computation, memory, and model components across distributed resources to improve scalability and efficiency. As a result, LLM […]