classification

This paper introduces ADS-C, an antidistillation defense for classification that provably preserves top-1 accuracy while degrading student model performance by up to 29.7 percentage points, achieving zero utility cost for the teacher.

Introduces 'Diffusing Blame', a method for neural networks to learn competitively while strictly enforcing Dale's principle, using error diffusion and modulo error routing, achieving strong results on image classification and reinforcement learning without weight transport.

This paper introduces Influence Matching (Inf-Match), a dataset distillation method that aligns the final training outcome by learning a compact synthetic set whose effect on converged parameters matches that of the full dataset. It achieves state-of-the-art accuracy on classification benchmarks and outperforms strong baselines on vision-language distillation tasks.

Proposes Parallel Quantum Feature Augmentation (PQFA), a hybrid quantum-classical framework that applies shallow variational quantum circuits to enhance fused multimodal features, outperforming classical baselines with fewer parameters.

This study empirically compares Kolmogorov-Arnold Networks (KANs) and Multi-Layer Perceptrons (MLPs) on structured tabular classification tasks, finding that KANs statistically outperform MLPs but with higher computational cost.

LMU researchers developed a Quantum Boltzmann Machine model for pneumonia detection from chest X-rays. Using fewer than 9,000 parameters (vs. millions in classical CNNs), the model achieves 84–86% accuracy, demonstrating potential advantages of quantum machine learning for small medical datasets.

The paper proposes a lightweight method that reformulates regression-based INR training as a classification task by discretizing continuous targets into bins, enabling flexible distribution modeling for error-aware uncertainty estimation in scientific data compression.

Proposes HDE-Net, a manifold-constrained deep neural network that uses hyperbolic space to better model rule-based structures in tabular data, achieving state-of-the-art performance on the TALENT-tiny-core benchmark while maintaining efficiency.

This paper introduces the TrustX Agent Risk Classification Framework (ARC), a structured instrument for risk-tiering internally created agentic AI systems, grounded in existing governance frameworks and featuring a twelve-dimension scoring rubric, autonomy levels, and three-tier governance output.

This article delves into transfer learning from the perspective of category theory, proposing deep manifold theory. It argues that neural networks learn relational structures through attribute-free numerical computation, thereby enabling cross-domain transfer, and explains the internal logic of classification.

Zer0Fit provides an MCP server that wraps Google's TabFM and TimesFM foundation models for zero-shot forecasting, classification, and regression tasks, running entirely locally.

普林斯顿博士后Shilong Liu提出自进化代理的三层分类体系：工件迭代优化、Agent Harness自我改进和无黄金答案的模型学习，系统梳理了相关概念和前沿工作。

GitHub released the Multilingual Repositories Dataset, covering over 40 million repositories and 80 million classification rows, with insights into non-English READMEs, issues, and PRs. Korean leads in issue text while Portuguese tops READMEs.

This paper studies the fairness problem of thresholded subgroup underdiagnosis in long-tailed chest X-ray classification, demonstrating that rare-label fairness depends jointly on the finding, subgroup, and operating threshold, not on label frequency or ranking metrics alone.

This paper investigates knowledge distillation for time series classification using FCN, Inception, and ConvTran architectures on the UCR Archive, showing significant parameter reduction and performance maintenance.

The paper proposes UASPL, a method that integrates predictive reliability into sample selection for self-paced learning using evidential neural networks, improving classification performance and interpretability.

This paper proposes the Intuitionistic Fuzzy Graph Embedded Random Vector Functional Link with Multiview Learning (IFGRVFL-MV) model, which integrates intuitionistic fuzzy sets, graph embedding, and multiview learning to improve classification accuracy and robustness to outliers. Experiments on benchmark datasets show that IFGRVFL-MV outperforms existing models.

This paper proposes replacing the proprietary GPT-3.5 metadata extractor in Multi-Meta-RAG with a lightweight, deterministic probe trained on hidden states of a small open-source model. The probe achieves 90.9% accuracy, outperforming GPT-3.5 (80.9%) and a substring baseline (88.0%), while avoiding allow-list drift and API costs.

This paper consolidates the state of the art in supervised political scaling, investigating whether joint prediction of ideological scales and a middle ground between classification and regression can improve performance.

Google Research introduces TabFM, a zero-shot foundation model for tabular data that uses in-context learning to perform classification and regression without requiring manual model training or hyperparameter tuning.