Evolve Quick Reviews
Public Paper Reviews
Structured preliminary reviews of recent open-access papers, generated and continuously updated by the Evolve analysis engine. Each review condenses what the paper claims, how solid the evidence is, and the questions a rigorous peer reviewer would ask — English analysis with abstracts in your language. These are independent AI analyses, not peer review; the canonical record always lives at the publisher.
TriWorldBench: A Tri-View Consistency Perspective on Embodied World Models
Xuanyi Liu, Haofeng Wang, Ruiqi Li, Danni Yu et al.
TriWorldBench: a tri-view embodied world-model benchmark (500 episodes, 19 metrics, TWB-Score); protocol is thoughtful but the extracted text contains no results or baselines.
BOBA: Dynamic Bayesian Optimization through Bayesian Active Inference
Merlin Angel Kelly, Rishan Patel, Alexander Thomas, Ziyue Zhu et al.
BOBA derives an active-inference acquisition function for dynamic BO; results are synthetic-only, mixed in time-limited runs, and use oracle priors plus per-cell tuned beta.
FIRE: Failure-Informed Runtime Engineering for Reliable Language-Model Agents
Nikita Agarwal, Nivedit Jain
Claims runtime policies from observed failures lift agent pass^2 on Terminal-Bench 2.1 across 3 GPT-5.6 tiers; rigorous design, honest stats, small N.
Compiling Sufficient Governance Context from Declared Losses and Reachable States: Exact Observation-Contract Synthesis with Cardinality and Cost Objectives
Gaston Besanson
Standard rough-set reduct synthesis via SAT/MaxSAT applied to constructed governance domains; rigorous and honest, but novelty modest and evidence synthetic.
The Dynamics of Quasiregular Neural Learning
Matthia Sabatelli
A clean toy study: neural nets acquire, lose, then recover exceptions during training; rarer exceptions overregularize more; mechanism untested.
Truth for Believable AI: Expressed Doubt, Provenance, and Belief Revision as an Engineerable Stance
Sebastian Cochinescu
An honesty behavior layer (doubt, provenance, revision) over a fixed 0.5B LLM; several pre-specified checks fail, and key gains rest on just 60 facts.
REVE: Efficient Hallucination Correction for Large Audio-Language Models via Reused Encoder States
Hongjin Song, Jiasheng Kuang, Xinyu Yang, Qiuyu Fang et al.
Proposes reusing LALM encoder states to verify audio event mentions; removes 92.9% of unsupported ones at 0.756 recall, but test sets (n=500) are small.
On the security and privacy of LLMs in Mobility
Mauro Conti, Lorenzo Perinello, Umberto Salviati
Claims first security survey of 35 LLM-mobility works (AI Act rubric); timely, but count contradictions and an untested 'significant' claim weaken it.
xWhyL: Causal Interactive Learning
Nicholas Tagliapietra, Florian Peter Busch, Moritz Willig, Matej Zečević et al.
Formalizes learning causal models from expert explanations (EEC, Tug-of-War theorem; CIL method); promising, but code pending and validation synthetic-only.
FairMean: Promoting Fairness in Distributed Learning under Label Poisoning Attacks
Huigan Zheng, Jiaojiao Zhang, Yongxiang Liu
FairMean's bounded loss-reweighting trades fairness vs poisoning robustness with clean theory and small-scale experiments; modest but sound contribution.
Canonical locks that encode part-whole hierarchies
Rajat Modi, Yogesh Singh Rawat
Proposes canonical locks: part-whole hierarchies encoded as relative phase angles in GLOM-style nets. Rich concepts; empirical evidence remains preliminary.
Learning to Fluctuate: Statistical Foundations for Causal Tabular Pretraining
Zhiheng Zhang
Fluctuation labels teach causal tabular PFNs a reusable sampling response; rigorous Lean-verified theory, but coverage depends on an accurate variance head.
Target alignment, dilution and forecast selection when cross-sectional forecasts share a common target
Masoud Soleimani
Paper shows standardized common-target forecasts split into alignment plus orthogonal parts; careful geometry, but empirics: selection removes dilution yet nothing beats no-info.
MAVP: Map-Aware Visuomotor Policies for Mobile Manipulation
Jinhe Tang, Ruixiao Dai, Weiming Zhi
Map-frame pose targets with localisation feedback lift success on 6 real mobile-manipulation tasks and 3 policies; strong ablations, no significance tests.
Blaming Across the Aisle: Political Contrasting and Blame Attribution in the Danish Parliament
Markus Lundsfryd Jensen, Rune Egeskov Trust, Kenneth Christian Enevoldsen, Sara Kolding
Claims Danish parliamentary blame fell to 2016 then rose, amplified by right-wing extremity; credible but silver-label circularity and n=13 caveats temper it.
EADC: Evaluation of Advanced and Deep-level Compliance in Large Language Models
Yan Zhang, Ruien Li, Yaoyao Peng, Wanxin Ren et al.
Law-grounded compliance benchmark (4,435 QA, 24 LLMs) is well-designed, but an unspecified judge LLM and no error bars weaken its human-baseline claims.
PreGS: A Parameter-Transfer-Based Multi-Expert Graph Neural Network for Node Classification
Zhicong Cai, Yinglong Zhang, Xiaoying Hong, Xuewen Xia et al.
Claims GAT→GraphSAGE weight-transfer multi-expert fusion; competitive gains on 8 graphs, but trivial theory, no significance tests, ensemble fairness unclear.
CompKV: Compensation-Aware KV Selection for Long-Context LLM Inference
Zhen Huang, Ruizhe Yao, Danyi Liu, Xinrui Chen et al.
Claims first compensation-aware KV selection; best sparse accuracy on RULER, up to 6.85x self-attention speedup; per-layer latency only, no significance tests.
HySparse2: Hybrid Sparse Attention with Two-Level KV Sharing
Jianyu Wei, Yizhao Gao, Qihao Zhang, Shimao Chen et al.
Proposes two-level KV sharing sparse attention that cuts prefill and KV cache with better retrieval; solid, though MQA switch confounds comparisons.
Layout-Guided Masking for GROBID: Lightweight Structural Gains in Large-Scale Scientific PDF Ingestion
Luca Foppiano, Sana Khamassi, Vipul Gupta
GROBID + 轻量级 CPU 区域检测器提升了结构化 PDF 解析效果（MS 段落召回率 +0.086，表格 F1 ≈0.93–0.98），评估严谨；收益源于 token 遮蔽，且存在披露的局限性。
Learning to Defer with Guidance on Real World Medical Data
Emma Sun, Joshua Strong, Alison Noble
First real-data test of L2D and a new AI-guidance routing setup; beats baselines on chest X-rays, but small real data and simulated guidance temper the claim.
COPE: Continual Personalization of LLMs under Sparse User Feedback via User Embeddings and Self-Evaluation
Ruike Cao, Fugen Yao, Liang Dong, Jian Xu et al.
COPE uses per-user soft embeddings and self-calibrated proxy rewards for continual personalization; gains look real but simulator-coupled; scores stay modest.
Error Bounds for Statistical Estimators in BTL Model with Parametric Multivariate Utility Functions
Yicheng Li, Huifu Xu
Extends BTL MLE theory to arbitrary fixed query designs with minimax bounds and existence threshold; rigorous, but constants lag the empirical regime.
Designing and Analysing Argument Mining Pipelines: Towards a Comprehensive Assessment
Siddharth Bhargava, Sara Tonelli, Patricia Martín-Rodilla
Preliminary meta-study of end-to-end argument mining pipelines via a triple-perspective framework; promising taxonomy, but analysis stays qualitative.