The paper presents a method that uses distributed acoustic sensing (DAS) and deep learning to monitor urban traffic at fine spatiotemporal resolution. By repurposing underground fiber‑optic cables as dense sensor arrays, the approach captures roadway activity at meter‑level spatial and second‑level temporal scales. A deep learning framework processes raw vibration waveforms to detect vehicle trajectories and infer traffic volume and speed, with a hybrid training strategy that combines synthetic and manually annotated data to improve detection in noisy, congested conditions.
By Hao Tian, Heng Cai, Xiaowei Chen, Yifan Yang
The paper introduces HermNet, a spectral graph neural network that uses Hermite polynomials for nodewise prediction and normalized propagation, avoiding eigendecomposition or learned bases. It compares HermNet to other complete polynomial bases, examining how coordinate choices affect optimization under limited training. Experiments on synthetic and real data show regimes where HermNet outperforms alternatives, and analyze the impact of calibration, regularization, and training duration on performance.
By Shuang Wu
SPADE-DFL is a communication‑efficient decentralized federated learning algorithm that uses a primal–dual method to allow the number of local function‑value updates between neighbor exchanges to increase with the computation budget while maintaining non‑private convergence rates. For smooth nonconvex objectives, it achieves a time‑averaged stationarity and consensus bound of ≠O(T−1/3) with only ≠Theta(T−2/3) communication rounds, where T is the number of local updates per client. The method also supports client‑level differential privacy by isolating data‑dependent increments, proving privacy for the full interactive transcript and quantifying the resulting optimization error, and demonstrates higher mean test accuracy than existing decentralized learning methods on four classification tasks.
By Mengli Wei, Mengkai Zhu, Jiawen Chen, Wenwu Yu, Duxin Che
arXiv:2609. 29458v1 Announce Type: new Abstract: We present a tightened convergence analysis of clipped gradient descent on $(L_0, L_1)$-smooth functions, with quantitative constants.
By David A. R. Robin
The paper formalizes slate recommendation as a two-stage process: a randomized score learner followed by deterministic selection. It shows that a differential‑privacy guarantee can be preserved through the selection stage when selector inputs are public or independent, and introduces a logged margin certificate that bounds score‑induced objective changes, ensuring the slate order remains stable. Experiments on OULAD, MovieLens‑25M, and Amazon Musical Instruments demonstrate that higher anchor weight reduces ranking churn, while simulations confirm bounded target drift and utility dependent on settings.
By Sam Urmian, Qinyi Liu, Mohammad Khalil
Synthetic Hospital is an open, fully synthetic longitudinal electronic health record benchmark created from public medical education material. It contains 1,268 patients and 5,602 encounters, with every diagnosis, finding, and temporal relation grounded in standard ontologies and traceable back to its source. Physicians could not reliably distinguish its records from real charts, and current frontier AI models perform significantly below human experts on tasks such as reconstructing problem lists and summarizing charts.
By Christine Park, Valerie Chen, Tim Dettmers
The paper introduces a new oversampling technique that uses a predictive model to generate nominal variables instead of interpolating them, addressing a key limitation of SMOTE. Applied to a large GPS‑location dataset of university students, the method outperforms existing oversampling approaches in predicting high levels of amotivation and anhedonia. This improvement could help identify students with these debilitating symptoms earlier and more accurately.
By Dang Nguyen, Bao Duong, Arun Kumar, Dat Phan-Trong, Julian Berk, Taylor Braund, Kien Do, Debopriyo Bal, Wu Yi Zheng, Leonard Hoon, Jill Newby, Helen Christensen, Svetha Venkatesh, Alexis Whitton, Sunil Gupta
The paper introduces a safety‑oriented pedestrian trajectory prediction framework for urban intersections that fuses pedestrian motion history with Time‑to‑Collision (TTC) data and crossing‑zone context. Using the inD dataset, a pooled LSTM architecture encodes TTC and context separately before integrating them with pedestrian positions, and a weighted loss emphasizes large errors. The approach reduces average displacement error (ADE) and final displacement error (FDE) and significantly lowers the frequency of errors exceeding a 1 m tolerance compared to a position‑only model.
By Erel Avineri, Yftach Gil, Yehudit Aperstein
The paper investigates why vision‑language models that tokenize images with vector‑quantized (VQ) codebooks frequently hallucinate objects on grounded yes/no tasks. By applying activation patching across 25 models from eight large‑language‑model families, the authors uncover an early‑layer attention routing circuit shared by VQ‑tokenized VLMs. They develop a three‑gate diagnostic that isolates ten models carrying this circuit, show that swapping a single architectural component (VQ+Linear) introduces the circuit, and demonstrate that ablating the early‑layer ($L_0$) component reduces hallucinations in open‑ended generation by 31 % while other decoding‑time fixes do not.
By Shamanthak Hegde, Xiangrui Liu, Maitreya Patel, Yezhou Yang
Sufficiently Reduced Distributional Regression (SRDR) is a generative approach that merges conditional distribution estimation with nonlinear sufficient dimension reduction (SDR). By framing SDR as a risk minimization problem using strictly proper scoring rules, SRDR jointly learns a dimension reduction map and a generative prediction model through minimization of the energy score, which can be estimated via sampling. The method extends to multi‑environment data and classification, and theoretical results show convergence of estimated conditional distributions in energy distance, implying asymptotic sufficiency. In experiments on CT slice localization, superconductivity data, and digit classification, SRDR recovers low‑dimensional sufficient structure and matches or surpasses state‑of‑the‑art nonlinear SDR methods in representation quality and predictive performance.
By Alexander Henzi, Tiange Liu, Xinwei Shen
The paper introduces Flow-Based Distribution Matching (FBDM), a non‑adversarial framework that learns self‑supervised representations using explicit geometric references and spherical conditional velocity regression. FBDM assigns augmented image views to shared target references while limiting reference usage, and employs an alignment loss to bring view representations closer. Experiments on datasets from CIFAR to ImageNet demonstrate that FBDM performs nearly as well as adversarial DM, outperforms existing SSL methods, and achieves a 1.48‑ to 1.83‑fold speedup with minimal GPU memory increase, while a theoretical analysis bounds downstream misclassification rates in terms of the pretraining loss.
By Yuling Jiao, Wensen Ma, Houduo Qi, Defeng Sun
The paper introduces a method for exact unlearning of Gibbs supervised learning algorithms via a variational formulation based on empirical risk minimization with relative entropy regularization (ERM‑RER). By maximizing the expected empirical risk over the data to be removed while regularizing with relative entropy to the original algorithm, the resulting solution is a new Gibbs probability measure that matches the distribution of an algorithm retrained from scratch on the remaining data. The approach also provides a general framework for reweighting data points in ERM‑RER, allowing for up‑ or down‑weighting to control generalization error or other objectives.
By Yaiza Bermudez, Samir M. Perlaza, I\~naki Esnaola
VG‑TIE is a tabular‑to‑image encoding method that uses Natural and Horizontal Visibility Graphs to map feature values into a two‑dimensional image via PCA. Each pixel represents a feature, its intensity shows deviation from the population mean, and edges encode visibility relationships. The method offers model‑agnostic, intrinsically interpretable images, providing feature ranking through node degree distributions and local/global importance via pixel intensity combined with Grad‑CAM, and demonstrates competitive performance on six public datasets.
By David Chushig-Muzo, Luis M. L\'opez-Ramos, \'Angeles Rodr\'iguez de Cara, Eva Milara, Luis Zhinin-Vera, Diego H. Peluffo-Ord\'o\~nez
The paper introduces $A^2E^2$, a diagnostic framework that decomposes confusion in fine‑grained aircraft detection into four distinct, measurable sources: affinity, heterogeneity, contested, and collapsed. By attributing each source to either aleatoric or epistemic uncertainty and within‑ or between‑class distinctions, the method provides actionable remedies and experimentally verifies that targeted interventions reduce the identified source without affecting irreducible ones. This turns passive confusion matrices into a concrete, validatable diagnosis that can guide model improvement.
By Hai Huang, Helmut Mayer
The paper investigates whether layer-wise visual‑text similarity in multimodal large language models (MLLMs) truly reflects content‑level cross‑modal interaction. By injecting Gaussian noise into the visual stream of 13 MLLMs, the authors show that task accuracy drops sharply while traditional scalar alignment metrics (CKA, SVCCA, MIR, principal‑angle cosine) fail to distinguish corrupted from clean inputs, a phenomenon they term the alignment illusion. They propose the principal‑angle gap (PA gap) as a more reliable geometric diagnostic that correlates better with task performance and reveals when internal geometry diverges from accuracy.
By Hong-Han Wang, Yuntao Wang, Hu Ding
The paper proposes a two‑stage framework, SL+LHF, that first learns low‑dimensional representations from noisy labeled data and then refines model alignment using human comparison feedback via a probabilistic bisection approach. It introduces the label‑noise‑to‑comparison‑accuracy (LNCA) ratio to theoretically identify when this framework outperforms pure supervised learning, showing that trading labels for comparisons reduces sample complexity when labels are scarce. Experiments on a high‑dimensional crowdfunding prediction task and an Amazon Mechanical Turk study confirm that incorporating human or large language model evaluators improves accuracy under a fixed query budget.
By Junyu Cao, Mohsen Bayati
MLPerf Automotive is the first standardized public performance benchmark for evaluating machine learning systems used in automotive AI acceleration. Developed by MLCommons, it addresses the unique constraints of automotive workloads—such as sensor suites, safety, and real‑time processing—that existing benchmarks cannot handle. The benchmark covers perception tasks (2D/3D object detection, semantic segmentation), end‑to‑end driving, and infotainment, and includes curated models, methodology, and reference implementations available on GitHub.
By Radoyeh Shojaei, Predrag Djurdjevic, Mostafa El-Khamy, James Goel, Kasper Mecklenburg, P{\i}nar Muyan-\"Oz\c{c}elik, John Owens, Tom St. John, Jinho Suh, Arjun Suresh
GOMA (Graph-Optimized Multimodal Alignment) introduces a dual-embedding approach for multimodal retrieval, separating content embeddings supervised for paired identity from semantic embeddings trained with cross-modal pairs and observed relationships. The method fuses these embeddings, applies semantic agreement to weight graph edges, and uses restart graph propagation to reinforce the initial signal, enabling both single-modality and dual-attribute retrieval. Across six datasets and four tasks, GOMA outperforms 14 external methods on 14 primary metrics, with controlled experiments highlighting the impact of separate supervision, graph regularization, and semantic-guided propagation.
By Xu Wang, Xunkai Li, Yinlin Zhu, Rong-Hua Li, Guoren Wang
The paper introduces MultiwayPAM, a tensor clustering technique designed for analyzing LLM-as-a-Judge score tensors that capture evaluations across questions, answerers, and evaluators. By simultaneously estimating cluster memberships and medoids for each mode, the method reveals the structure of bias inherent in large language model evaluators. Experiments on two practical datasets demonstrate the effectiveness of MultiwayPAM in uncovering these bias patterns.
By Chihiro Watanabe, Jingyu Sun
The paper investigates how post‑training compression techniques—such as pruning, quantization, and distillation—affect demographic fairness in Whisper speech‑recognition models. It finds that pruning and INT4 quantization significantly widen word‑error‑rate gaps between demographic groups, especially for Black/AA and Asian speakers, while distillation tends to reduce these gaps. The study introduces a temporal‑taxation metric to quantify the increased correction effort required for marginalized speakers after compression.
By Srishti Ginjala, Eric Fosler-Lussier, Christopher W. Myers, Srinivasan Parthasarathy