paper-with-me

홈 › Papers

A Logic-Driven Framework for Consistency of Neural Models

2019-08-31 · IJCNLP 2019 11 · Tao Li, Vivek Gupta, Maitrey Mehta, Vivek Srikumar

While neural models show remarkable accuracy on individual predictions, their internal beliefs can be inconsistent across examples. In this paper, we formalize such inconsistency as a generalization of prediction error. We propose a learning framework for constraining models using logic rules to regularize them away from inconsistency. Our framework can leverage both labeled and unlabeled examples and is directly compatible with off-the-shelf learning schemes without model redesign. We instantiate our framework on natural language inference, where experiments show that enforcing invariants stated in logic can help make the predictions of neural models both accurate and consistent.

📄 PDF Abstract BibTeX arXiv:1909.00126

Code (1)

utahnlp/consistency 공식 구현 pytorch

Tasks

Natural Language Inference

Similar Papers 제목 키워드 기반

Aligning with Logic: Measuring, Evaluating and Improving Logical Preference Consistency in Large Language Models

2024-10-03 · Yinhong Liu, Zhijiang Guo, Tianya Liang, Ehsan Shareghi 외

Large Language Models (LLMs) are expected to be predictable and trustworthy to support reliable decision-making systems. Yet current LLMs often show inconsistencies in their judgments. In this work, we examine logical pr…

Decision MakingNegation

A neurosymbolic Approach with Epistemic Deep Learning for Hierarchical Image Classification

2026-05-11 · Ezel Kilicdere, Shireen Kudukkil Manchingal, Fabio Cuzzolin arxiv

Deep neural networks achieve high accuracy on image classification tasks. Yet, they often produce overconfident predictions as which fail to express epistemic uncertainty, and frequently violate logical or structural con…

Image Classification

Ontology-Driven Structural Regularization for Document-Level Relation Extraction

2026-08-21 · Laura Menotti, Stefano Marchesin, Gianmaria Silvello arxiv

Document-Level Relation Extraction (DocRE) relies heavily on costly manually annotated datasets, while large distant supervision resources such as DocRED distant remain underexploited due to noise. We show that a critica…

Document-level Relation Extraction

GeoReason: Aligning Thinking And Answering In Remote Sensing Vision-Language Models Via Logical Consistency Reinforcement Learning

2026-01-07 · Wenshuai Li, Xiantai Xiang, Zixiao Wen, Guangyao Zhou 외 arxiv

The evolution of Remote Sensing Vision-Language Models(RS-VLMs) emphasizes the importance of transitioning from perception-centric recognition toward high-level deductive reasoning to enhance cognitive reliability in com…

Reinforcement Learning

Instance-Aware Robust Consistency Regularization for Semi-Supervised Nuclei Instance Segmentation

2025-10-10 · Zenan Lin, Wei Li, Jintao Chen, Zihao Wu 외 arxiv

Nuclei instance segmentation in pathological images is crucial for downstream tasks such as tumor microenvironment analysis. However, the high cost and scarcity of annotated data limit the applicability of fully supervis…

Instance Segmentation