arXiv:2605.06341v1 Announce Type: cross Abstract: Many real-world optimization problems consist of multiple tightly coupled subproblems whose solutions must be coordinated to achieve high overall performance. However, existing large language model driven automated heuristic design approaches are limited to single-problem settings. In this paper, we propose CoupleEvo. CoupleEvo proposes three evolutionary coordination strategies to evolve heuristics […]
How to make the most of your masked language model for protein engineering
arXiv:2603.10302v2 Announce Type: replace-cross Abstract: A plethora of protein language models have been released in recent years. Yet comparatively little work has addressed how to best sample from them to optimize desired biological properties. We fill this gap by proposing a flexible, effective sampling method for masked language models (MLMs), and by systematically evaluating models […]
ANCORA: Learning to Question via Manifold-Anchored Self-Play for Verifiable Reasoning
arXiv:2604.27644v2 Announce Type: replace-cross Abstract: We propose a paradigm shift toward open-ended curriculum self-play: rather than learning to answer on a fixed prompt set, a unified policy learns to question: generating verifiable problems, solving them, and turning verifier feedback into self-improvement without human-annotated solutions. We introduce ANCORA, in which the policy alternates between a Proposer […]
A $mu$-distance for semidirected orchard phylogenetic networks
arXiv:2605.06243v1 Announce Type: cross Abstract: In evolutionary biology, phylogenetic networks are now widely used to represent the historical relationships between species and population, when this history includes reticulation events such as hybridization, gene flow and admixture between populations. Semidirected phylogenetic networks are appropriate models when the direction of some edges and the root position are […]
Activation in Vesicle-Mediated Signaling Shaped by Batch Arrival Statistics
arXiv:2605.06456v1 Announce Type: cross Abstract: Vesicle-mediated secretion of ions or molecules is a central mechanism of cellular communication, for example in processes such as neurotransmission or hormone release. These events are inherently stochastic: vesicle fusions lead to bursts of variable sizes, releasing discrete packets of transmitters that are subsequently cleared or degraded. The dynamics break […]
UniPool: A Globally Shared Expert Pool for Mixture-of-Experts
arXiv:2605.06665v1 Announce Type: cross Abstract: Modern Mixture-of-Experts (MoE) architectures allocate expert capacity through a rigid per-layer rule: each transformer layer owns a separate expert set. This convention couples depth scaling with linear expert-parameter growth and assumes that every layer needs isolated expert capacity. However, recent analyses and our routing probe challenge this allocation rule: replacing […]
Beyond Factual Correctness: Mitigating Preference-Inconsistent Explanations in Explainable Recommendation
arXiv:2603.03080v2 Announce Type: replace Abstract: LLM-based explainable recommenders can produce fluent explanations that are factually correct, yet still justify items using attributes that conflict with a user’s historical preferences. Such preference-inconsistent explanations yield logically valid but unconvincing reasoning and are largely missed by standard hallucination or faithfulness metrics. We formalize this failure mode and propose […]
DeTrigger: A Gradient-Centric Approach to Backdoor Attack Mitigation in Federated Learning
arXiv:2411.12220v3 Announce Type: replace-cross Abstract: Federated Learning (FL) enables collaborative model training across distributed devices while preserving local data privacy, making it ideal for mobile and embedded systems. However, the decentralized nature of FL also opens vulnerabilities to model poisoning attacks, particularly backdoor attacks, where adversaries implant trigger patterns to manipulate model predictions. In this […]
RobustSora: De-Watermarked Benchmark for Robust AI-Generated Video Detection
arXiv:2512.10248v2 Announce Type: replace-cross Abstract: The proliferation of AI-generated video models poses new challenges to information integrity and digital trust. A key confound, however, remains unaddressed: commercial generators embed visible overlay watermarks for provenance tracking, yet no existing benchmark controls for this variable, leaving open whether detectors learn genuine generation artefacts or merely associate watermark […]
Multimodal Fact-Level Attribution for Verifiable Reasoning
arXiv:2602.11509v2 Announce Type: replace-cross Abstract: Multimodal large language models (MLLMs) are increasingly used for real-world tasks involving multi-step reasoning and long-form generation, where reliability requires grounding model outputs in heterogeneous input sources and verifying individual factual claims. However, existing multimodal grounding benchmarks and evaluation methods focus on simplified, observation-based scenarios or limited modalities and fail […]
Screening Is Enough
arXiv:2604.01178v3 Announce Type: replace-cross Abstract: A core limitation of standard softmax attention is that it does not provide an independently interpretable measure of query–key relevance: attention scores are unbounded, while attention weights are defined only relative to competing keys. Consequently, irrelevant keys cannot be explicitly rejected, and some attention mass is assigned even when no […]
Attributions All the Way Down? The Metagame of Interpretability
arXiv:2605.06295v1 Announce Type: cross Abstract: We introduce the metagame, a conceptual framework for quantifying second-order interaction effects of model explanations. For any first-order attribution $phi(f)$ explaining a model $f$, we measure the directional influence of feature $j$ on the attribution of feature $i$, denoted as meta-attribution $varphi_j to i(f)$, by treating the attribution method itself […]