This paper evaluates a method to improve representations using incomplete external evidence across tasks.
problem Increasing labelled data quality and quantity is challenging due to manual labelling errors and noise.
method Evidence Transfer method using incomplete categorical external evidence.
result Evidence Transfer proves effective and robust against different levels of incompleteness.
There is growing evidence that converting targets to soft targets in supervised learning can provide considerable gains in performance. Much of this work has considered classification, converting hard zero-one values to soft labels---such as by adding label noise, incorporating label ambiguity or using distillation. In…
Martingale Doppelgänger-Eval benchmarks VLMs on candlestick evidence vs. trend extrapolation
problem Auditing whether VLMs use chart evidence or trend extrapolation
method Proving formal limitations and designing controlled mechanisms
result Identifying regression coefficients for evidence vs. trend
Paper presents a Siamese network for identifying more convincing evidence.
problem Identifying the more convincing argument in discussions.
method Proposes a Siamese neural network architecture for labeling evidence as more convincing.
result The Siamese network outperforms baselines on a new challenging data set.
In the recent years, we have witnessed the development of multi-label classification methods which utilize the structure of the label space in a divide and conquer approach to improve classification performance and allow large data sets to be classified efficiently. Yet most of the available data sets have been provide…
Crowdsourcing has become very popular among the machine learning community as a way to obtain labels that allow a ground truth to be estimated for a given dataset. In most of the approaches that use crowdsourced labels, annotators are asked to provide, for each presented instance, a single class label. Such a request c…
We present a new boosting algorithm, motivated by the large margins theory for boosting. We give experimental evidence that the new algorithm is significantly more robust against label noise than existing boosting algorithm.
Label smoothing improves generalization by controlling generalization loss.
problem Lack of mathematical understanding of label smoothing's effectiveness.
method Proposed a theoretical framework to show how label smoothing controls generalization loss in the label noise setting.
result Predicted an optimal label smoothing point that minimizes generalization loss.
Interpolating label noise makes models vulnerable to adversarial attacks.
problem Adversarial vulnerability of models trained on noisy labels.
method Theoretical analysis of label noise and adversarial risk relationship.
result Uniform label noise induces adversarial risk similar to worst-case poisoning.
This paper improves neural tangent kernels for better generalization and local elasticity.
problem Performance gap between neural tangent kernels and real-world neural networks.
method Introduces label-aware kernels using Hoeffding decomposition.
result Models trained with proposed kernels simulate NNs better in terms of generalization and local elasticity.
A new method improves uncertainty estimation in deep learning, especially for hard-to-label samples.
problem Improving uncertainty estimation for hard-to-label samples in deep learning.
method Introduces Fisher Information Matrix (FIM) to dynamically reweight objective loss terms.
result Consistently outperforms traditional evidential neural networks in uncertainty estimation tasks.
New f-divergence measures improve robustness in noisy label learning.
problem Improving robustness in learning with noisy labels.
method Derived decoupling property of f-divergence measures under label noise. result Properly defined f-divergence measures are robust with label noise. DNNs improve accuracy by using more evidence from images.
problem Understanding why DNNs generalize well and improving model selection metrics.
method Minimal sufficient views (MSVs) to identify key evidence regions in images.
result DNNs with more evidence regions in images have higher generalization performance.
We derive and analyze a new, efficient, pool-based active learning algorithm for halfspaces, called ALuMA. Most previous algorithms show exponential improvement in the label complexity assuming that the distribution over the instance space is close to uniform. This assumption rarely holds in practical applications. Ins…
Credit expansion led to stronger household leverage cycles during the U.S. business cycle.
problem Understanding the role of credit supply in the U.S. business cycle.
method Causal evidence from 1999-2010 U.S. business cycle data.
result Credit expansion, particularly in private-label mortgages, caused stronger household leverage cycles.
Fairness audits fail under missing protected labels, especially at zero access.
problem Understanding the reliability of fairness audits with incomplete protected-label data.
method Introduced a seed-calibrated stress test to separate missingness effects from seed-to-seed movement.
result Missing protected labels do not significantly alter fairness mitigation methods, but they can lead to harmful intersectional outcomes.
A new algorithm detects changepoints in labeled and unlabeled data.
problem Accurate detection of abrupt changes in partially labeled data.
method Labeled Optimal Partitioning (LOPART) algorithm that fits train labels and predicts unlabeled changepoints.
result LOPART provides more accurate predictions than existing methods in both train and test sets.
This paper identifies and estimates the label noise transition matrix without ground truth labels.
problem Learning with noisy labels and identifying the noise transition matrix.
method Building on Kruskal's identifiability results, the paper characterizes the identifiability of the label noise transition matrix for the generic case at the instance level.
result The necessity of multiple noisy labels in identifying the noise transition matrix for the generic case at the instance level.
Improves neural networks by adjusting labels to reduce overfitting to noisy data.
problem Reduces overfitting to mislabeled or non-representative samples in neural networks.
method Combines self-adaptive training with mixup to improve accuracy and robustness.
result Achieves state-of-the-art accuracy on image recognition datasets with label noise.
This work explores how neural network architecture affects robustness to noisy labels.
problem The impact of neural network architecture on robustness to noisy labels.
method Formal framework connecting robustness to architecture alignments, measured by predictive power in representations.
result Network robustness to noisy labels improves when its architecture is more aligned with the target function.
Fine-grained pretraining improves neural network's ability to learn rare features.
problem Improving generalization in deep learning models.
method Introducing a hierarchical multi-view structure to confine input data distribution.
result Fine-grained pretraining leads to better accuracy on hard downstream test samples.
Paper tackles instance-dependent label noise by approximating it with part-dependent noise.
problem Learning with instance-dependent label noise is challenging.
method Approximate instance-dependent label noise with part-dependent noise. Use transition matrices for parts to model noise.
result Method outperforms state-of-the-art approaches for instance-dependent label noise.
New method pools labels from similar data items to improve learning from small samples.
problem Learning from small, human-annotated samples with potential disagreement among annotators.
method Proposes neighborhood-based pooling for sharing labels across similar data items.
result Improves learning from small, noisy samples by pooling labels from similar items.
We offer a novel view of AdaBoost in a statistical setting. We propose a Bayesian model for binary classification in which label noise is modeled hierarchically. Using variational inference to optimize a dynamic evidence lower bound, we derive a new boosting-like algorithm called VIBoost. We show its close connections …
This paper refines human labeling as a measurement process, revealing four sources of variation.
problem Systematic variation in human labeling obscures model learning.
method Introduces a statistical framework to decompose labeling outcomes.
result Empirical evidence for four components of labeling variation.
We create a large dataset for fact checking claims and improve prediction accuracy.
problem Fact checking claims from multiple sources is challenging.
method We created a comprehensive dataset and developed a novel method for automatic veracity prediction.
result Our model achieves a Macro F1 of 49.2%, showing significant performance improvements.
Adding uninformative labels improves tumor segmentation in low-data mammography.
problem Improving tumor segmentation in mammography with limited data.
method Used seemingly uninformative labels from non-expert annotators to turn a multi-label task into a multi-class problem.
result Performance gains in tumor segmentation are achieved in low-data settings with additional uninformative labels.
Prevents sensitive data generation in diffusion models using labeled and unlabeled data.
problem Generating sensitive data in diffusion models using unlabeled data.
method Positive-Unlabeled Diffusion Models, approximating ELBO with labeled and unlabeled data.
result Prevents the generation of sensitive data without compromising image quality.
This work presents a new strategy for multi-class classification that requires no class-specific labels, but instead leverages pairwise similarity between examples, which is a weaker form of annotation. The proposed method, meta classification learning, optimizes a binary classifier for pairwise similarity prediction a…
We propose Rademacher complexity bounds for multiclass classifiers trained with a two-step semi-supervised model. In the first step, the algorithm partitions the partially labeled data and then identifies dense clusters containing κ predominant classes using the labeled training examples such that the proportion of t…
Enhanced nearest neighbor improves accuracy in crowdsourced data.
problem Noise in crowdsourced labels degrades classification accuracy.
method Developed two algorithms to estimate worker quality and an enhanced nearest neighbor classifier.
result Proposed methods achieve the same regret as oracle version based on expert data.
Detecting and recovering labels in binomial logistic mixtures is challenging due to an information gap.
problem Detecting and recovering labels in binomial logistic mixtures
method Propose two feasibility-aware inference procedures
result Avoid misleading component selections and improve label probability calibration
Proposes a game-theoretic approach for class-dependent rationalization.
problem Optimizing feature selection for complex neural predictors.
method A game-theoretic approach where classes compete to find evidence for factual and counterfactual scenarios.
result The method identifies both factual and counterfactual rationales consistent with human rationalization.
We consider two questions at the heart of machine learning; how can we predict if a minimum will generalize to the test set, and why does stochastic gradient descent find minima that generalize well? Our work responds to Zhang et al. (2016), who showed deep neural networks can easily memorize randomly labeled training …
Study shows cross-domain X-ray prediction performance discrepancies and label shifts.
problem Quantifying generalization limits across different X-ray datasets.
method Large-scale study on multiple X-ray datasets, focusing on performance and label shifts.
result Interesting discrepancies found between model performance and agreement, and concept similarity across tasks.
Given a constant magnetic field on Euclidean space Rp determined by a skew-symmetric (p×p) matrix Θ, and a Zp-invariant probability measure μ on the disorder set Σ which is by hypothesis a Cantor set, where the action is assumed to be minimal, the corresponding Integrated Density…
Supervised object detection and semantic segmentation require object or even pixel level annotations. When there exist image level labels only, it is challenging for weakly supervised algorithms to achieve accurate predictions. The accuracy achieved by top weakly supervised algorithms is still significantly lower than …
Develops active learning method for linear optimization with margin-based criterion.
problem Optimizing decisions in linear optimization problems with limited labeled data.
method Smart Predict-then-Optimize (SPO) loss and margin-based active learning algorithm.
result Algorithm achieves significantly fewer labels than naive supervised learning, especially for minimizing SPO loss.
New method detects TC imagery patterns for rapid intensity change.
problem Detecting upcoming rapid intensity changes in TC satellite imagery.
method Nonparametric test of association between images and event labels using neural networks and bootstrap.
result Identifies archetypes of infrared imagery associated with elevated rapid intensification risk.
Unified VAE framework improves unsupervised, semi-supervised, and supervised learning.
problem Improving performance in learning tasks with limited labeled data.
method A VAE with a classification layer connected to the encoder and combined with the latent layer for the decoder, supplemented with a supervised loss for labeled data.
result The approach outperforms direct supervised setups and boosts unsupervised tasks with unlabeled data.
Dual-edge spatial Jacobian image graph for interpretable diabetic retinopathy grading
problem Automated diabetic retinopathy grading from color fundus photographs
method Dual-edge spatial-Jacobian image graph
result 0.8076 accuracy, 0.8312 quadratic weighted kappa, 0.5915 macro-F1, 0.9330 adjacent-grade accuracy
In conventional supervised pattern recognition tasks, model selection is typically accomplished by minimizing the classification error rate on a set of so-called development data, subject to ground-truth labeling by human experts or some other means. In the context of speech processing systems and other large-scale pra…
Method detects batch heterogeneity in genomic data.
problem Batch effects confound genomic diagnostics.
method Bayesian model evidence clustering.
result Detects batch effects without known labels.
New model shows SGD can prefer sharp or flat solutions based on label noise.
problem Understanding SGD's preference for flat or sharp solutions during training.
method Solved an analytically solvable model to explore SGD behavior.
result Data distribution determines sharpness at convergence; isotropic label noise leads to flat minimum preference.
Deep networks can overfit benignly but still be vulnerable to adversarial attacks.
problem Adversarial vulnerability of deep neural networks trained with benign overfitting.
method Investigated causes of adversarial vulnerability, identified label noise as a key factor, and explored the impact of training procedures and representation learning.
result Adversarial robustness requires more complex decision boundaries than simple ones, suggesting the need for better representation learning.
Paper presents a multi-label topic model for financial texts with high performance and insights into market reactions.
problem Analyzing financial text data for market reactions and understanding topic interactions.
method Trained a multi-label topic model on a financial text database, achieved high macro F1 score, and investigated topic interactions.
result Model achieves high performance (macro F1 > 85%) and reveals significant market reactions to topic co-occurrences.
Study uses LLMs to categorize financial tweets, revealing useful sentiment signals.
problem Discovering meaningful sentiment signals from unstructured financial social media data.
method Leveraged LLMs to automatically label financial tweets with event categories and aligned with returns.
result Certain event labels consistently yield negative alpha, with statistically significant Sharpe ratios and information coefficients.
Modern computer vision algorithms typically require expensive data acquisition and accurate manual labeling. In this work, we instead leverage the recent progress in computer graphics to generate fully labeled, dynamic, and photo-realistic proxy virtual worlds. We propose an efficient real-to-virtual world cloning meth…