Exploring Imbalanced Annotations for Effective In-Context Learning

• URL: http://arxiv.org/abs/2502.04037v1
• Date: Thu, 06 Feb 2025 12:57:50 GMT
• Title: Exploring Imbalanced Annotations for Effective In-Context Learning
• Authors: Hongfu Gao, Feipeng Zhang, Hao Zeng, Deyu Meng, Bingyi Jing, Hongxin Wei,
• Abstract summary: We show that imbalanced class distributions in annotated datasets significantly degrade the performance of in-context learning (ICL)<n>Our method is motivated by decomposing the distributional differences between annotated and test datasets into two-component weights.<n>Our approach can prevent selecting too many demonstrations from a single class while preserving the effectiveness of the original selection methods.
• Score: 41.618125904839424
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Large language models (LLMs) have shown impressive performance on downstream tasks through in-context learning (ICL), which heavily relies on the demonstrations selected from annotated datasets. Existing selection methods may hinge on the distribution of annotated datasets, which can often be long-tailed in real-world scenarios. In this work, we show that imbalanced class distributions in annotated datasets significantly degrade the performance of ICL across various tasks and selection methods. Moreover, traditional rebalance methods fail to ameliorate the issue of class imbalance in ICL. Our method is motivated by decomposing the distributional differences between annotated and test datasets into two-component weights: class-wise weights and conditional bias. The key idea behind our method is to estimate the conditional bias by minimizing the empirical error on a balanced validation dataset and to employ the two-component weights to modify the original scoring functions during selection. Our approach can prevent selecting too many demonstrations from a single class while preserving the effectiveness of the original selection methods. Extensive experiments demonstrate the effectiveness of our method, improving the average accuracy by up to 5.46 on common benchmarks with imbalanced datasets.

Related papers

• Revisiting Demonstration Selection Strategies in In-Context Learning [66.11652803887284]
  Large language models (LLMs) have shown an impressive ability to perform a wide range of tasks using in-context learning (ICL) In this work, we first revisit the factors contributing to this variance from both data and model aspects, and find that the choice of demonstration is both data- and model-dependent. We propose a data- and model-dependent demonstration selection method, textbfTopK + ConE, based on the assumption that textitthe performance of a demonstration positively correlates with its contribution to the model's understanding of the test samples.
  arXiv  Detail & Related papers  (2024-01-22T16:25:27Z)
• Revisiting Long-tailed Image Classification: Survey and Benchmarks with New Evaluation Metrics [88.39382177059747]
  A corpus of metrics is designed for measuring the accuracy, robustness, and bounds of algorithms for learning with long-tailed distribution. Based on our benchmarks, we re-evaluate the performance of existing methods on CIFAR10 and CIFAR100 datasets.
  arXiv  Detail & Related papers  (2023-02-03T02:40:54Z)
• CLIPood: Generalizing CLIP to Out-of-Distributions [73.86353105017076]
  Contrastive language-image pre-training (CLIP) models have shown impressive zero-shot ability, but the further adaptation of CLIP on downstream tasks undesirably degrades OOD performances. We propose CLIPood, a fine-tuning method that can adapt CLIP models to OOD situations where both domain shifts and open classes may occur on unseen test data. Experiments on diverse datasets with different OOD scenarios show that CLIPood consistently outperforms existing generalization techniques.
  arXiv  Detail & Related papers  (2023-02-02T04:27:54Z)
• Adaptive Dimension Reduction and Variational Inference for Transductive Few-Shot Classification [2.922007656878633]
  We propose a new clustering method based on Variational Bayesian inference, further improved by Adaptive Dimension Reduction. Our proposed method significantly improves accuracy in the realistic unbalanced transductive setting on various Few-Shot benchmarks.
  arXiv  Detail & Related papers  (2022-09-18T10:29:02Z)
• Learning to Re-weight Examples with Optimal Transport for Imbalanced Classification [74.62203971625173]
  Imbalanced data pose challenges for deep learning based classification models. One of the most widely-used approaches for tackling imbalanced data is re-weighting. We propose a novel re-weighting method based on optimal transport (OT) from a distributional point of view.
  arXiv  Detail & Related papers  (2022-08-05T01:23:54Z)
• Learning to Adapt Classifier for Imbalanced Semi-supervised Learning [38.434729550279116]
  Pseudo-labeling has proven to be a promising semi-supervised learning (SSL) paradigm. Existing pseudo-labeling methods commonly assume that the class distributions of training data are balanced. In this work, we investigate pseudo-labeling under imbalanced semi-supervised setups.
  arXiv  Detail & Related papers  (2022-07-28T02:15:47Z)
• BASIL: Balanced Active Semi-supervised Learning for Class Imbalanced Datasets [14.739359755029353]
  Current semi-supervised learning (SSL) methods assume a balance between the number of data points available for each class in both the labeled and the unlabeled data sets. We propose BASIL, a novel algorithm that optimize the submodular mutual information (SMI) functions in a per-class fashion to gradually select a balanced dataset in an active learning loop.
  arXiv  Detail & Related papers  (2022-03-10T21:34:08Z)
• CMW-Net: Learning a Class-Aware Sample Weighting Mapping for Robust Deep Learning [55.733193075728096]
  Modern deep neural networks can easily overfit to biased training data containing corrupted labels or class imbalance. Sample re-weighting methods are popularly used to alleviate this data bias issue. We propose a meta-model capable of adaptively learning an explicit weighting scheme directly from data.
  arXiv  Detail & Related papers  (2022-02-11T13:49:51Z)
• Minority Class Oriented Active Learning for Imbalanced Datasets [6.009262446889319]
  We introduce a new active learning method which is designed for imbalanced datasets. It favors samples likely to be in minority classes so as to reduce the imbalance of the labeled subset. We also compare two training schemes for active learning.
  arXiv  Detail & Related papers  (2022-02-01T13:13:41Z)
• Relieving Long-tailed Instance Segmentation via Pairwise Class Balance [85.53585498649252]
  Long-tailed instance segmentation is a challenging task due to the extreme imbalance of training samples among classes. It causes severe biases of the head classes (with majority samples) against the tailed ones. We propose a novel Pairwise Class Balance (PCB) method, built upon a confusion matrix which is updated during training to accumulate the ongoing prediction preferences.
  arXiv  Detail & Related papers  (2022-01-08T07:48:36Z)
• BALanCe: Deep Bayesian Active Learning via Equivalence Class Annealing [7.9107076476763885]
  BALanCe is a deep active learning framework that mitigates the effect of uncertainty estimates. Batch-BALanCe is a generalization of the sequential algorithm to the batched setting. We show that Batch-BALanCe achieves state-of-the-art performance on several benchmark datasets for active learning.
  arXiv  Detail & Related papers  (2021-12-27T15:38:27Z)
• Prototypical Classifier for Robust Class-Imbalanced Learning [64.96088324684683]
  We propose textitPrototypical, which does not require fitting additional parameters given the embedding network. Prototypical produces balanced and comparable predictions for all classes even though the training set is class-imbalanced. We test our method on CIFAR-10LT, CIFAR-100LT and Webvision datasets, observing that Prototypical obtains substaintial improvements compared with state of the arts.
  arXiv  Detail & Related papers  (2021-10-22T01:55:01Z)
• PLM: Partial Label Masking for Imbalanced Multi-label Classification [59.68444804243782]
  Neural networks trained on real-world datasets with long-tailed label distributions are biased towards frequent classes and perform poorly on infrequent classes. We propose a method, Partial Label Masking (PLM), which utilizes this ratio during training. Our method achieves strong performance when compared to existing methods on both multi-label (MultiMNIST and MSCOCO) and single-label (imbalanced CIFAR-10 and CIFAR-100) image classification datasets.
  arXiv  Detail & Related papers  (2021-05-22T18:07:56Z)
• Attentional-Biased Stochastic Gradient Descent [74.49926199036481]
  We present a provable method (named ABSGD) for addressing the data imbalance or label noise problem in deep learning. Our method is a simple modification to momentum SGD where we assign an individual importance weight to each sample in the mini-batch. ABSGD is flexible enough to combine with other robust losses without any additional cost.
  arXiv  Detail & Related papers  (2020-12-13T03:41:52Z)
• Sequential Targeting: an incremental learning approach for data imbalance in text classification [7.455546102930911]
  Methods to handle imbalanced datasets are crucial for alleviating distributional skews. We propose a novel training method, Sequential Targeting(ST), independent of the effectiveness of the representation method. We demonstrate the effectiveness of our method through experiments on simulated benchmark datasets (IMDB) and data collected from NAVER.
  arXiv  Detail & Related papers  (2020-11-20T04:54:00Z)
• Classification Performance Metric for Imbalance Data Based on Recall and Selectivity Normalized in Class Labels [0.0]
  We introduce a new performance measure based on the harmonic mean of Recall and Selectivity normalized in class labels. This paper shows that the proposed performance measure has the right properties for the imbalanced dataset.
  arXiv  Detail & Related papers  (2020-06-23T20:38:48Z)
• Improving Multi-Turn Response Selection Models with Complementary Last-Utterance Selection by Instance Weighting [84.9716460244444]
  We consider utilizing the underlying correlation in the data resource itself to derive different kinds of supervision signals. We conduct extensive experiments in two public datasets and obtain significant improvement in both datasets.
  arXiv  Detail & Related papers  (2020-02-18T06:29:01Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.