←── back to feed
/topics/arxiv-computer-science-papers-september-25-part-2

arXiv computer science papers September 25 part 2

16 items●1 sources●updated 44d ago●trend 0

Sixteen machine learning papers posted to arXiv on September 25, 2026 span interpretability, drug discovery, diffusion models, time-series forecasting, and domain-specific applications including thyroid cytology classification, hydrogen embrittlement detection, and molecular diversity selection.

  • SMILESGNN combines SMILES Transformer and GATv2 graph encoders for interpretable drug toxicity prediction with class-imbalance handling.
  • VINTAGE-TS adapts time-series foundation models to account for data revisions, distinguishing observation time from information-availability time.
  • TAM-Chain uses Absorbing Markov Chains and Shannon Entropy for multi-scale thyroid cytology classification with false-negative suppression.
  • CARE introduces condition-aware representation regularization as a plug-and-play framework for diffusion model training efficiency.
  • fable.intermittent R package benchmarks probabilistic forecasting methods for intermittent demand across unified software framework.
[BLG]blog/rss16
Stable and Faithful Explanations for Knowledge Tracing
arXiv cs.LG · Praveena Padi, Arun Morampudi, Ujval Sai Gopal Irrinki, Pradeep Kumar Dolabehera Kakitapelli · 44d
SMILESGNN: Interpretable Clinical Toxicity Prediction via SMILES-Graph Cross-Attention Fusion
arXiv cs.LG · Quang Minh Nguyen, Thuy Quynh Nguyen, Duc Minh Le, Ho Nhat Minh Nguyen, Thanh Long Dai Doan, Trong Nghia Nguyen · 44d
CFD Correction of Open Tip Clearance Flow in a Compressor Cascade Using VAE Latent Space Adaptation
arXiv cs.LG · Xiang Zuo, Hefang Deng, Caiyan Chen, Honglin He, Mingmin Zhu, Songan Zhang, Jinfang Teng · 44d
CARE: Condition-Aware Representation Regularization for Diffusion Models
arXiv cs.LG · Fengjia Guo, Zhuoyi Yang, Jie Tang · 44d
SpaFactor: Lightweight Spatial Context-Aware Gene Program Modeling for Histology-to-Transcriptomics Inference
arXiv cs.LG · Shiting Ruan, Xitong Ling, Qiming He, Ziyou Yan, Huaitian Yuan, Tian Guan, Ying Xiao, Xu Guan, Yonghong He · 44d
When Explanations Cannot Be Read: Measuring and Correcting SHAP and LIME Rendering for Right-to-Left Languages
arXiv cs.LG · Rameesha Zia, Muhammad Shahid Iqbal Malik · 44d
Leakage-Safe Machine Learning for Hydrogen Embrittlement Detection in 316L Stainless Steel: A Region-Held-Out Evaluation of Texture and Deep Features in SEM Micrographs
arXiv cs.LG · Muhammad Awais, Muhammad Yaseen, Abdul Shakoor, Niaz Ahmed Niaz, Huria Zia, Muhammad Zain Shakoor · 44d
Time-Series Foundation Models That Understand Data Revisions
arXiv cs.LG · Taimoor Ahmad · 44d
Uncovering Residential PV-EV Co-Adoption from Smart-Meter Data: Load Archetypes and Detection for Demand-Side Planning
arXiv cs.LG · Jack Zheng, Hao Wang · 44d
Auditability Is Not One Property: Rule Overlap, Behavioural Agreement, and Composition in Reinforcement Learning
arXiv cs.LG · Liu Hung Ming · 44d
TAM-Chain: Multi-Scale Thyroid Cytology Classification via Absorbing Markov Chains and Shannon Entropy Uncertainty Quantification for False-Negative Suppression and Domain-Shift Adaptation
arXiv cs.LG · Hai Pham Ngoc · 44d
Learning to Discover Interesting Mathematics
arXiv cs.LG · Niket Patel, Ahmad Rammal, Amaury Hayat, Remi Munos, Julia Kempe · 44d
Physics-Informed Self-Supervised Learning for Joint Wire Calibration and Interaction Position Reconstruction in Multi-Wire Parallel Plate Avalanche Counters
arXiv cs.LG · Antoine Lemasson, Maurycy Rejmund · 44d
UO-FIE: Combining Exact-Label Supervision with Graded Utility for Factivity Inference
arXiv cs.LG · Xinchen Xiao · 44d
fable.intermittent: benchmarking probabilistic forecasting methods for intermittent time series
arXiv cs.LG · Stefano Damato, Lorenzo Zambon, Giorgio Corani, Dario Azzimonti · 44d
OPDiv: Optimal Selection of Top-K High-Scoring, Diverse Compounds
arXiv cs.LG · Miroslav L\v{z}i\v{c}a\v{r} (Deep MedChem) · 44d