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.
Adaptive Agent Design
Raj Kiriti Velicheti, Subhonmesh Bose, Tamer Başar
Proposes bi-level design of an agent's internal-state kernel; proves soft Q-learning convergence plus ZOO/BO outer-loop rates; theory sound, experiments thin.
GraphProfiler: Source-Linked Sensitive Attribute Inference via Personal Knowledge Graphs
Ahmed Sohair Khan, Estrid He, Chenglong Ma, Monica Wachowicz et al.
Claims auditable LLM attribute inference via source-linked PKGs at near-baseline ASR; evidence validation strong, but graph-vs-RAG ablation absent, stats thin.
Bridging Vision Foundation Model Priors with CLIP for Spatial-aware Few-shot Anomaly Detection in Medical Images
Juzheng Miao, Yuchen Yuan, Cheng Chen, Pheng-Ann Heng
Spatial-FAD injects DINO spatial priors into CLIP for few-shot medical AD; Dice gains are large (avg +11.4%, 4-shot) but AUC is mixed and variance unreported.
TripPattern: A Pattern-based Text Watermarking Method for Large Language Models
Sangjun Moon, Dasom Choi, Jingun Kwon, Hidetaka Kamigaito et al.
Three-group alternation watermarking that preserves LLM text quality better than KGW; broad multilingual results, but weak against paraphrase attacks.
Information Specialization and Constrained Synthesis in Multi-Agent LLM Forecasting: A Prospective Live-Study of the 2026 FIFA World Cup
Julian Varghese, Lucas Bickmann, Sarah Sandmann
A prospective 56-match live study finds news-specialized LLM forecasts edge quantitative ones, but n=56 descriptive differences are too small to conclude.
EvoRS: On-Policy Self-Evolution of Reward Systems for Open-Ended Reinforcement Learning
Weiyuan Li, Aili Chen, Xintao Wang, Yikai Zhang et al.
EvoRS evolves the full reward system as an executable DAG during RL; strong roleplay gains (+4.767 CoSER), but the writing margin over RaR sits within noise.
Linear Exponential Quadratic Gaussian Covariance Steering
Chiran B. Cherian, Yasemin Isik, Abhishek Halder
Theory letter extends LQG covariance steering to the risk-sensitive LEQG objective; rigorous local existence-uniqueness proof, thin single-instance numerics.
Beyond Vector Similarity: Hierarchical Context-Aware Graph RAG vs Standard RAG in Enterprise Code Migration
Nilesh Jaiswal, Aniket Agrawal, Arjit Shukla, Divya Malhotra et al.
Graph RAG cut API hallucinations (56.4%→16.2%) vs Standard RAG on one repo, but single-codebase evidence and inconsistent composite scores limit the claim.
Not All Speech Is Intent: Adaptive Self-Correcting Inference Layer for Post-ASR False Wake-Up
Preeti Saraswat, Divya Neelagiri, Anil Yadav
Samsung ASCIL: post-ASR feedback loop cuts false wake-ups (54.27% on failure-only subset, 24.39% issue-tagged); feedback heuristics unvalidated.
A Differentially Private Federated Proximal Optimization Framework for Customer Churn Prediction in Heterogeneous Federated Telecom Networks
Joydeb Kumar Sana, Subrata Chakraborty, M M Manjurul Islam
Claims first DP-FedProx for telecom churn; results plausible but gains are sub-pp, single-seed, untested for significance, and DP-FedAvg often wins.
AMDKernelVault: Large-Scale Datasets and Agentic Training for AMD GPU Kernel Optimization
Ji Liu, Saptarshi Majumder, Yiqing Huang, Wenwen Ouyang et al.
AMD's execution-verified HIP/Triton kernel corpus trains an 8B model that tops correctness on 3 benchmarks; evals are small and semantic overlap unaudited.
When Does AI Augment Work? A Workflow-Level Framework for Human-Agent Collaboration
CIVIC-AI Collaboration, :, Jiaying Wu, Caleb Ziems et al.
Workshop whitepaper defines six conditions for genuine AI augmentation via a workflow-level lens; conceptually solid, empirically untested framework.
$\text{GSF-}χ$: Global Stereochemical Fields for Chiral Graph Transformers
Jiaqing Xie, Yuxin Wang, Xipeng Qiu
GSF-χ is a chiral graph transformer with proven mirror-parity laws; axial gains are strong (+12.6% Rotation) but rest on small test sets and n=3 runs.
Zipbench: Low-Cost Framework for Compressing Comprehensive Benchmarks of Large Language Models
Zhongzhan Huang, Junxin Li, Guoming Ling, Yupei Lin et al.
Claims low-cost compression of 100+ LLM benchmarks from 6 anchors (~0.98 Spearman); solid theory, but small held-out sets and OOD error need scrutiny.
Computing at Sea: Floating and Offshore Data Centres as a Pathway to Sustainable AI Infrastructure
Cheng Siong Chin, Jianhua Zhang, M. Venkateshkumar
Claims offshore data centres cut cooling energy and water use; timely, readable synthesis, but no original analysis and the land-side baseline is dodged.
Confidence-Gated Transductive Test Generation for Code Reranking
Sungjae Lee, Youngsik Yoon, Seockbean Song, Siwei Wang et al.
CoTT gates transductive test generation by inductive confidence, beating prior baselines on both benchmarks, though ungated ablation wins on HumanEval-R+.
Poisson-Corrector Complexity Bounds for Moreau--Yosida Unadjusted Langevin Sampling
Yuchen Xin, Zhihua Zhang
MYULA invariant bias O(h)+Õ(h^{3/4}) via Poisson corrector, giving Õ(ε^{-4/3}) complexity; rigorous and novel, but key proofs were absent from the excerpt.
A Splitting Method for SDE Terminal-Law Estimation
Rushil Gupta, Sandeep Juneja
Strong CLT + minimax PAVA allocation theory for SDE path splitting; 10-25% error gains hold, but tuned constant-splitting often rivals the learned allocation.
Temporal Recurrence Favors Fewer Layers
Ivan Anokhin, Johan Obando-Ceron, Irina Rish, Sebastian Risi
Claims recurrence lets streaming models match deep nets with fewer layers; sweeps are solid, but LM results lack seeds and throughput gains unproven.
Learning-Augmented Optimization for Strategic Two-Echelon Spare Parts Network Design
Donato Maragno, Marco Caserta, Alberto Sinigaglia, Komlanvi Ametana et al.
Claims 30.5% savings vs exact baseline in Amazon spare-parts design via GNN-guided search; strong protocol but single proprietary case, no significance tests.
The House with a Million Windows: Interactive Fiction for Narrative Restorying
Cody Kommers, Sarah G Immel, Drew Hemment, Mina Lee
LLM interactive-fiction tool for restorying personal stories; NISE gains (dz=0.67, N=40) lack a control group—engaging design, causal evidence still weak.
Optimizing Geoengineering Interventions Using Differentiable Climate Models
Pulkit Dubey, Dorian S. Abbot, Ashesh Chattopadhyay
Backprop through differentiable GCMs designs SST cooling removing 92.3% of land warming; rigorous but idealized: prescribed SST, no ocean or aerosol physics.
RoofLang: Enabling AI-Driven Architecting of LLM Inference Systems
Ziyue Yang, Yuting Jiang, Lei Qu, Peng Cheng
Bold DSL+simulator for AI-driven LLM inference architecting; results are uncalibrated simulation and a promised grid-search comparison is absent.
Calibrated Ambiguity in Multimodal Language Models: Humans reach for cultural references, while models describe the picture
Cody Kommers, Mingrui Ye, Evelyn Gius, Daniela Mihai et al.
Claims open-weight VLMs show ambiguity collapse and cultural flattening on Dixit vs humans; well-validated rubric but design confound and no frontier models.