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.
On the SoS Certifiability of Log-Concave Distributions
Aleksandr Storozhenko
Claims optimal O(m) SoS moment certificates for all log-concave distributions via stochastic localization; proof chain verifies but rests on unrefereed preprints.
SciWalker: Synthesizing Scientific Coding Problems with Operator Graphs and Execution Feedback
Chenxi Li, Wenxuan Zeng, Yun Luo, Fangchen Yu et al.
Claims operator-graph synthesis of 8,178 scientific coding problems; strong RL gains but no ablations, no error bars, checkpoint picked on SciCode test.
Style, Not Self: Surface Cues Explain Zero-Shot Code Attribution by Large Language Models
Ehsan Barkhordar, Surendrabikram Thapa
Rigorous negative study: zero-shot LLM code self-recognition tracks length/style cues, not self-knowledge; caveats: mostly small models, short MBPP code.
NNV3: Expanding Neural Network Verification to New Architectures and Domains
Anne M. Tumlin, Samuel Sasaki, Ben Wooding, Diego Manzanas Lopez et al.
NNV3 unifies weight-perturbation, video, GNN, probabilistic and fairness verification in one MATLAB tool; integration is credible but evaluations stay small.
Era by Eon: Benchmarking Enterprise Agents on Hidden Knowledge
Benjamin Gruenbaum, Doron Porat, Assaf Natanzon, Roy Zavida et al.
Adds 8 deterministic hidden-knowledge questions to Era by Eon; elegant construct, honest limits, but single-seed eval and internal count errors weaken claims.
Can Labor Markets Function in the Age of AI? The Evaluation Bottleneck in Hiring
Itai Ashlagi, Ramesh Johari, Jon Kleinberg, Anushka Murthy
Theory: AI makes application signals noisy, so firms screen by experience and may exclude high-fit entrants; intermediate assessments restore screening.
Scoring Both Directions: LLMs realize the MRS they cannot reliably parse
Soham Dan
Claims prompted LLMs match trained realizers on MRS→text but fail parsing (65.5 vs 91.0 EDM); well-controlled, though stats and model coverage are thin.
Self-Play Pretraining with Zero Data
Aditya Cowsik, Kfir Dolev, Michael Y. Li, G. Bruno De Luca et al.
Zero-data self-play over a UTM program space shows predictable zero-shot scaling on natural data; small-scale but well-controlled proof of concept.
A Native-Reference Phone-Class Geometry for Second-Language Pronunciation Analysis
Tina Raissi, Nhan Phan, Chenxiao Wang, Mikko Kurimo
The paper proposes an interpretable phone-class geometry whose distances negatively correlate with L2 proficiency (ρ −0.53/−0.34); solid, honestly caveated, but much of the signal overlaps basic fluency features.
How Reproducible Are Evaluation Conclusions? A Self-Audit of LLM-Inferred Prompt Structure
Dipankar Sarkar
A self-audit showing LLM prompt-structure rankings are statistically fragile; unusually honest, well-instrumented, but rests on tiny samples from one endpoint.
Nuclear Norm-Regularized Bayesian Matrix Completion
Calvin Tolbert
首个针对核范数矩阵补全的、有保证的多项式时间贝叶斯采样器；理论可信，但边界仅具可行性，实验规模较小。
Residual Correlation as a Diagnostic for Joint-Uncertainty Gains from GP Coregionalisation
Fangqin Zhou, Joaquin Vanschoren
Claims residual correlation, not target correlation, predicts when GP coregionalisation helps joint NLL; 16-benchmark evidence is strong, scope delimited.
JevOut: Natural Context Can Flip Decision Models
Zixiang Xu
Claims short natural context flips decision models (TFR 61-73%); rigorous adaptive attack study, though white-box access and auto answer checks temper it.
Exposome factors predict the association between sleep health and depression
Liu, W., Kuppers, V., Bi, H., Mahdipour, M. et al.
Exposome factors moderately predict a UKB sleep–depression phenotype (r=0.25, R²=0.06); rigorous validation, modest yield, intervention claims overreach.
Reproducing Transparent and Scrutable Recommendations: Exploring Open-Weight Models via Natural-Language User Profiles
Noah Mamié, Laurin van den Bergh
Rigorous UPR reproducibility confirms accuracy claims under test-set reranking, but its own steering probes undermine scrutability; the conclusion overstates.
A Mathematical Model of Motivated Emotional Mind - Cognitive Embodied System
Wiesław L. Galus, Janusz A. Starzyk
Formalizes the embodied, affect-driven MEM cognitive architecture via definitional equations; coherent but purely theoretical—no simulations, data, or proofs.
Return or Revise? Learning When Revision Helps Retrieval-Augmented QA
Nicholas Kashani Motlagh, Tim Anderson, Jeremy Gwinnup, Grant Erdmann
Paired revision grading yields consistent ranking gains over confidence but sub-point, DPR-only-significant gains; standard RAG beats revision as an alternative.
AT-SKM-Net: An Accelerated Trainable Sampling Kaczmarz-Motzkin Framework for Linear Hard-Constraint Feasibility on Dynamic Graphs
Xiaochen Zhang, Haoyu Zhu, Yao Zhang, Qingchun Hou
AT-SKM-Net speeds hard-constraint feasibility on dynamic graphs via GNN active-set sampling and Cholesky updates; strong gains, but speedup figures conflict.
PrivDrift: Auditing User-Secret Leakage Under Topic Drift in Active LLM Conversations
Luciano Maldonado
Shows LLM user secrets remain recoverable after topic drift (38.7–54.6%); controlled, statistically careful, but leakage framing lacks a suppression baseline.
TouchSight: Bare-Handed Tactile Prediction from Egocentric Video via Generative Visual Augmentation
Danyan Zhou, Jinxuan Lu, Jiawei Lin, Tianxing Chen et al.
Predicts bare-hand force from egocentric RGB by generatively re-rendering glove data; no force baselines and quantitative eval only in the AIGC domain.
R-DEIM Net: An Efficient Rationale-Augmented Dual-Expert Interaction Model for Paraphrase Detection
Pushp, Vaibhav Prajapati, Himangshu Sarma
Claims a 76M dual-expert model matches LLM QQP accuracy with rationales; promising, but cross-protocol comparisons and inconsistent ablation figures weaken it.
MQSS-Selector: RL-Guided Pass Selection for an MLIR Compilation Pipeline
Andre Youssefi, Ercüment Kaya, Minh Chung, Jorge Echavarria et al.
Promising RL pass-selection for MLIR quantum compilation (DAEP works; device ML solid), but unified scheduler unimplemented and baselines missing.
Multimodal Thinking with Renderable Programs
Sunli Chen, Ding Zhong, Ziqiao Ma, Jiaxin Liu et al.
SVGLM tunes VLMs to sketch SVG overlays while reasoning, gaining ~6.5-10 weighted pts over SFT on geometry; but no significance tests or release stated.
Screen Before You Serve: Simulation for Production Customer Experience AI Agents at 140M Scale
Edesio Alcoba, Kevin Rossell, Aman Gupta, Shao Tang et al.
Nubank's simulation screening tracks production agent rankings (r=0.74) and precedes live A/B wins; credible industry evidence, but tNPS gain conflates scope.