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.
Potential of Artificial Intelligence Algorithms for Identification of Relevant Diagnostic and Prognostic Biomarkers of Early-Stage Liver Cancer
Ali Bou Nassif, Darko Castven, Manar Abu Talib, Jibran Sualeh Muhammad et al.
Claims 90.74% five-stage HCC classification & in-vitro DNAJB14 validation; real translational value, but leakage and no per-class metrics weaken ML claims.
Human-Grounded Calibration for Long-Text Image-Text Congruence in Vision-Language Models
Alessandro Gambetti, Qiwei Han
Proposes CS, a Platt-style calibration for frozen VLM encoders; shows retrieval and calibration diverge, but human evidence rests on 25 items, one backbone.
ModaLens: Measuring Image Sensitivity in Report-Conditioned Medical VLMs
Sebastián Andrés Cajas Ordóñez, Maximin Lange, Quang Bui, Anqi Peter Li et al.
Paired image-swap audit shows VLMs answer from reports not images; rigorous and honest, but report-derived labels limit clinical conclusions.
Predictive Likelihood Ratios for Language Model Watermark Detection
Li Ma
Extends pivotal LLM watermark tests via Bayes-factor averaging over signal uncertainty; rigorous theory and evaluation, though a simple baseline often wins.
Empathy Is Steerable but Multi-Axial: Mechanism Geometry and Persona Effects in LLMs
JuHeon Ha, Byounghan Lee, Yunseo Choi, Kyung-Ah Sohn
Empathy ER/IP/EX vectors steer reliably at mid layers yet share structure; persona shifts mostly escape the basis — honest, but proxy-evaluator-bound evidence.
RAG-CT: Mitigating Privacy Risks on Retrieval-Augmented Generation Systems via Scanning Prompt Distribution
Xingyu Lyu, Jiayimei Wang, Jianfeng He, Ning Wang et al.
Claims entropy+margin query screening blocks RAG PII leakage; strong results but no false-positive/utility analysis and internal inconsistencies.
CIDERS: Cloud-Edge LLM Collaborative Learning via Accelerating Personalized Bilevel Optimization
Victor H. Chen, Hairui Yu, Stella K. Chung, Hong Yan
Claims first personalized bilevel cloud-edge LLM framework; theory clean but personalization evidence sits within noise and Table VI concedes 2x communication.
CiteShade: Citation Laundering in Multi-Source Retrieval-Augmented Generation and Its Counterfactual Defense
Guo Fuzheng
First RAG citation-laundering attack: wrong answers 0.01→0.68 with trusted-source citations; honest causal analysis but small-n, no-CI evaluation.
RESKILL: Explicit Failure Attribution and Structured Repair for Interactive Language Agents
Mengyi Deng, Xin Li, Duyi Pan, Zilin Wang et al.
Proposes structured, hypothesis-weighted skill repair for LLM agents; gains are consistent across six settings but small and lack statistical testing.
Coaching Qwen3 Coder 30B to Think Like a CodeClash Arena Agent
Ivy Ning Zhang
ReAct-style SFT reportedly lifts Qwen3-Coder-30B past its base and Qwen3 Coder Plus in CodeClash; promising but small-n, no significance tests, single arena.
Scalability and Performance Evaluation of Federated Learning Frameworks: A Comparative Analysis
Bassel Soudan, Sohail Abbas, Ahmed Kubba, Manar Wasif Abu Talib et al.
Claims OpenFL is most scalable of four FL frameworks on MNIST (2–100 clients); useful data, but unverifiable setup and inconsistent stats weaken trust.
New Conditions for Philosophers to Catch the Wave of Citizen Deliberation in the Age of Artificial Intelligence in advance
Bernard Reber
Philosophy paper proposing five democratic principles, 19 indicators, five AI roles for LLM-backed deliberation; coherent framework, zero empirical evaluation.
Predicting build orientation for SLM dental parts: a comparison of rotation representations and direct vector regression
Felix Schmalzel, Reimar Waitz, Moritz Kronberger, Thorsten Schöler
Compares 13 rotation representations ×2 backbones for SLM dental orientation; thorough and honest, but single runs, no baselines, private data limit strength.
More Than Just Access: Generative AI as Communication Intermediary for Blind and Low-Vision Users
Protik Dey, Mohd Saifuzzaman, Taslima Akter
Timely interview study (n=19) framing GenAI as a communication intermediary for BLV users; valuable framing, but thin evidence and limited novelty.
Beyond Accuracy: Robustness, Cost, and Governance Trade-offs for Vision-Language Models in Templated Document Extraction
Kushal Patel, Pushkal Shrivastava, Mackenzie Lees, Qirui Lu et al.
在合成支票数据集上，经过微调的开源 VLM 达到了 F1 0.985，而表现最好的商业模型为 0.93；这项研究严谨且诚实，但微调与零样本（FT-vs-zero-shot）对比的不对称性削弱了其核心结论的说服力。
Knowledge-Enriched Structured EHR Features for 30-Day Hospital Readmission Prediction on MIMIC-IV
Mohamad Najafi, Hongyun Fu, Mathias Brochhausen, Jian Wu et al.
Knowledge-enriched structured EHR features hit AUROC 0.743 (CV) on MIMIC-IV readmission, competitive with note-based models; demographics dominate the gain.
Solving Finite-sum Coupled Compositional Optimization via Multi-block-Single-probe Estimator
Wei Jiang, Sifan Yang, Yibo Wang, Lijun Zhang et al.
Journal extension proposing MSVR variance reduction for FCCO; strong theory and experiments, but only SOX baseline and no variance reporting.
When Does Test-Time Physical Diagnosis Pay? A Frozen Policy Buys Evidence It Never Reads
Zhengshu Zhang
Convincing occurrence proof that a frozen policy can buy full-trace evidence yet answer from memory; narrow single-pipeline scope, no code release.
Merging the Knowledge of LLMs for Automatic Speech Recognition
Hayato Futami, Tatsuya Kawahara
Task-arithmetic merging of domain LLM LoRAs into LLM-based ASR brings consistent modest gains with zero inference cost; sound work, significance untested.
Task-Directed Residual AddUNet:Perfect-Reconstruction Routing for Full-Rate Representations
Vikram R. Lakkavalli
AddUNet reframed as perfect-reconstruction routing: algebra correct but trivial; TIMIT PER gains (28.60→25.76) lack a capacity-matched non-PR baseline.
Per-Matrix Optimality Is Not Enough: Three-Level Optimization for Low-Rank LLM Compression
Huicheng Zhang, Xiyao Feng, Ze-Tong Li, Chengkai Zhu et al.
Extends SVD-LLM with block + full-model stages; large, honestly-hedged perplexity gains, but utility limited as only PPL is claimed.
LongAgent: History-Guided Agentic Search for Longitudinal Outcome Prediction
Siyao Wang, Florian Guitton, Shuojie Fu, Guanyu Tao et al.
LLM agent searches variable/window/aggregation combos for longitudinal prediction; synthetic gains are marginal (ΔRMSE 0.015, p=0.027) and clinical R²≈0.02.
Where Should the KV Cache Live? Placement Policies Across GPU, CPU, and SSD for Long-Lived Sessions
Srikanta Datta Tumkur, Jay Iyer, Mehar Simhadri, Sai Pavan Kumar et al.
Sim study: KV tiering's 73× session gain is pure capacity, not policy; predicted-reuse and prefetch recommendations fail; honest but limited to batch-one sim.
Calibrate, Then Route: A Measured Study of Learned Request Routing for Disaggregated LLM Serving
Srikanta Datta Tumkur, Jay Iyer, Mehar Simhadri, Sai Pavan Kumar et al.
Credible hardware study: calibrated learned router tops goodput on 8 A40s (0.864 vs 0.835-0.847), but small scale, 3 traces, no significance tests.