paper-with-me

Papers

Re-weighting Tokens: A Simple and Effective Active Learning Strategy for Named Entity Recognition

2023-11-02 · Haocheng Luo, Wei Tan, Ngoc Dang Nguyen, Lan Du

Active learning, a widely adopted technique for enhancing machine learning models in text and image classification tasks with limited annotation resources, has received relatively little attention in the domain of Named Entity Recognition (NER). The challenge of data imbalance in NER has hindered the effectiveness of active learning, as sequence labellers lack sufficient learning signals. To address these challenges, this paper presents a novel reweighting-based active learning strategy that assigns dynamic smoothed weights to individual tokens. This adaptable strategy is compatible with various token-level acquisition functions and contributes to the development of robust active learners. Experimental results on multiple corpora demonstrate the substantial performance improvement achieved by incorporating our re-weighting strategy into existing acquisition functions, validating its practical efficacy.

📄 PDF Abstract BibTeX arXiv:2311.00906

Code (0)

등록된 구현이 없습니다.

Tasks

Active Learningimage-classificationImage Classificationnamed-entity-recognitionNamed Entity RecognitionNamed Entity Recognition (NER)NER

Similar Papers 제목 키워드 기반

When Importance Sampling Misallocates Credit: Asymmetric Ratios for Outcome-Supervised RL

2025-10-07 · Jiakang Wang, Runze Liu, Qingpeng Cai, Lei Lin 외 arxiv

Reinforcement learning (RL) has shown great promise in large language models (LLMs) post-training, which typically rely on token-level clipping to maintain stability during optimization. Despite the empirical success of …

Reinforcement Learning

Logit Reweighting for Topic-Focused Summarization

2025-07-07 · Joschka Braun, Bálint Mucsányi, Seyed Ali Bahrainian arxiv

Generating abstractive summaries that adhere to a specific topic remains a significant challenge for language models. While standard approaches, such as fine-tuning, are resource-intensive, simpler methods like prompt en…

Prompt Engineering

Seeing the Image: Prioritizing Visual Correlation by Contrastive Alignment

2024-05-28 · Xin Xiao, Bohong Wu, Jiacong Wang, Chunyuan Li 외

Existing image-text modality alignment in Vision Language Models (VLMs) treats each text token equally in an autoregressive manner. Despite being simple and effective, this method results in sub-optimal cross-modal align…

cross-modal alignment

ReAttn: Improving Attention-based Re-ranking via Attention Re-weighting

2026-02-23 · Yuxing Tian, Fengran Mo, Weixu Zhang, Yiyan Qi 외 arxiv

The strong capabilities of recent Large Language Models (LLMs) have made them highly effective for zero-shot re-ranking task. Attention-based re-ranking methods, which derive relevance scores directly from attention weig…

Inverse-Dirichlet Weighting Enables Reliable Training of Physics Informed Neural Networks

2021-07-02 · Suryanarayana Maddu, Dominik Sturm, Christian L. Müller, Ivo F. Sbalzarini

We characterize and remedy a failure mode that may arise from multi-scale dynamics with scale imbalances during training of deep neural networks, such as Physics Informed Neural Networks (PINNs). PINNs are popular machin…