AdvCL flips adversarial perturbations from a threat into a geometry tool that cuts catastrophic forgetting and boosts robustness in continual LLM fine-tuning
Clinical vision-language models can re-link de-identified chest X-rays to their original radiology reports using cosine similarity alone — and a targeted privacy fix cuts the risk by 62% without hurting diagnostic utility.
SAGC dynamically adjusts GRPO group sizes during training to eliminate straggler bottlenecks, cutting wall-clock time nearly in half while matching or beating static baselines on reasoning benchmarks.
FLAME exploits a physics-level flaw in diffusion models to pinpoint AI-edited image regions at pixel precision, beating existing detectors across seven benchmarks
Researchers show that Industry 4.0 digital twin models already contain everything needed to auto-generate AI planning problems, no PDDL expertise required
AgentRedBench exposes how every major LLM fails indirect prompt injection attacks from SaaS tools, then ships a guard that cuts success rates from 70% to 2.4%
A new benchmark reveals that stitching multiple short video clips together reliably breaks safety guardrails in today's best multimodal models — and the fix may already be in your toolkit.
S3TS merges Monte Carlo Tree Search with stochastic scenario trees, letting energy planners handle nonlinear models and renewable uncertainty at the same time — something neither approach could do alone.
The Insertion Process learns where, what, and when to insert tokens, using variational inference over permutations to discover data-driven generation orders without left-to-right constraints.
VLBM uses variational latent decomposition to keep forecasts reliable during rare, high-impact distribution shifts — without sacrificing normal-case accuracy.
A new 18-language benchmark exposes a critical gap: LLMs can recognize figurative idioms but consistently fail when those same idioms are used literally
A dual-branch system combining a lightweight video encoder and a large VLM catches subtle procedural errors in first-person video, tackling the rare-mistake imbalance problem head-on
A new paper proves that long context windows in time series models aren't just about memory -- they're solving a hidden identification problem that changes how we should design forecasters.
Researchers from Nanjing University introduce TELBench and DRIFT to pinpoint exactly where deep-research agents go wrong mid-trajectory, not just whether they get the final answer right.