MMbeddings reduces categorical embeddings by treating them as latent effects, significantly decreasing parameters and mitigating overfitting.
problem Large cardinalities in categorical embeddings lead to high parameter counts and overfitting.
method MMbeddings treats embeddings as latent random effects in a variational autoencoder framework, reducing parameter count and mitigating overfitting.
result MMbeddings consistently outperforms traditional embeddings across various tasks, demonstrating its potential in machine learning applications.
The recently proposed unlabeled-unlabeled (UU) classification method allows us to train a binary classifier only from two unlabeled datasets with different class priors. Since this method is based on the empirical risk minimization, it works as if it is a supervised classification method, compatible with any model and …
This paper explores memorization in adversarial training and proposes a mitigation algorithm.
problem Understanding and mitigating robust overfitting in adversarial training.
method Demonstrated the capacity of deep networks to memorize adversarial examples, analyzed convergence and generalization issues, and proposed a new mitigation algorithm.
result Identified robust overfitting as a significant drawback of adversarial training and proposed a mitigation algorithm.
Framework mitigates overfitting in quantitative trading strategies.
problem Overfitting during strategy transition from backtest to live trading.
method Three-stage protocol: IS, WFA, OOS; majority pass, purge gaps, cliff veto, etc.
result Demonstrates how to detect overfitting through performance decay and drawdown behavior.
The paper investigates overfitting in hyperparameter optimization.
problem Overfitting in hyperparameter optimization (overtuning).
method Formal definition, large-scale reanalysis of HPO benchmark data, analysis of factors affecting overtuning.
result Overtuning is more common than previously assumed, leading to worse generalization error in 10% of cases.
A fast geometric regularizer improves event camera performance.
problem Event collapse in contrast maximization framework.
method Geometric regularizer to mitigate overfitting.
result State-of-the-art accuracy with reduced computational complexity.
New algorithm reduces overfitting in neural networks.
problem Overfitting in neural networks.
method Integrates SMC with SGHMC for mini-batch sampling.
result SMCSGHMC outperforms SGD and deep ensembles.
ARO overfits by making constraints dependent on uncertainty, leading to brittleness.
problem ARO's adaptive policies become brittle when realizations fall outside the uncertainty set.
method Assigning constraint-specific uncertainty set sizes with probabilistic guarantees.
result Regularization through specific uncertainty set sizes ensures stability and flexibility.
Efficient regularization mitigates catastrophic overfitting in single-step adversarial training.
problem Catastrophic overfitting in single-step adversarial training.
method ELLE regularization term to enforce local linearity of the loss function.
result Our regularization term effectively mitigates catastrophic overfitting without the drawbacks of previous methods.
New method regularizes deep networks by distilling self-knowledge.
problem Overfitting in deep neural networks.
method Self-knowledge distillation to regularize class-wise predictions.
result Significant improvement in generalization and calibration.
Label noise in adversarial training leads to robust overfitting, explained and mitigated.
problem Label noise in adversarial training causes robust overfitting.
method Proposed a method to automatically calibrate labels.
result Consistent performance improvements across various models and datasets.
UREs lead to overfitting in complex models, especially in complementary label learning.
problem Overfitting in weakly supervised learning with complementary labels.
method Proposed a surrogate complementary loss (SCL) framework to reduce gradient variance.
result SCL mitigates overfitting and improves URE-based methods.
Analyzes dropout in neural networks using differential equations.
problem Lack of theoretical understanding of dropout's effectiveness.
method Analytical study of dropout in two-layer neural networks with online stochastic gradient descent.
result Dropout reduces detrimental correlations and optimal probability increases with data noise.
Proposes a method to prevent overfitting in deep DRE models.
problem Overfitting in deep DRE models using empirical Bregman divergence.
method Introduces a non-negative correction for empirical Bregman divergence.
result The proposed method mitigates train-loss hacking and improves performance.
Deep networks can memorize random labels; symmetric loss improves this.
problem Deep networks can memorize random labels, ignoring standard regularization.
method Empirical studies with MNIST and CIFAR-10 datasets, formal definition of robustness.
result Symmetric loss function improves network's ability to resist memorization.
Bayesian Pseudo Label Selection reduces overfitting in semi-supervised learning.
problem Overfitting in pseudo-label selection for semi-supervised learning.
method BPLS, a Bayesian framework that approximates the posterior predictive of pseudo-samples.
result BPLS outperforms traditional PLS methods, especially in high-dimensional data.
The paper develops a statistical theory explaining overfitting in imbalanced classification.
problem Overfitting in high-dimensional imbalanced classification.
method Developed a statistical theory for support vector machines and logistic regression.
result Overfitting is more severe for the minority class due to truncation or skewing effects in high-dimensional data.
Optimal portfolios are formed by combining momentum, size, and volatility characteristics, enhancing utility for all investors.
problem Estimation error in forming optimal portfolios from characteristics.
method Maximizing an in-sample loss function that is more concave than the utility function, linking weights to characteristics.
result Optimal portfolios with significantly higher certainty equivalents than benchmarks for all investors.
Resetting from checkpoints improves DNN training with noisy labels.
problem Latent gradient bias induced by noisy labels causes overfitting.
method Stochastic resetting applied to SGD to mitigate latent gradient bias.
result Resetting significantly improves DNN generalization performance.
Adapts data analysis for growing data, improving generalization guarantees.
problem Challenges of overfitting and statistical validity in adaptive workflows with growing data.
method Generalizes adaptive analysis on dynamic data, incorporating time-varying empirical accuracy bounds and mechanisms.
result First generalization bounds for adaptive analysis on dynamic data, matching prior works' improvement over data splitting.
We propose self-adaptive training---a new training algorithm that dynamically corrects problematic training labels by model predictions without incurring extra computational cost---to improve generalization of deep learning for potentially corrupted training data. This problem is crucial towards robustly learning from …
FedCONST adapts update magnitudes to enhance feature generalization in FL.
problem Heterogeneous client data in FL leads to overfitting and distorted transferable features.
method FedCONST uses linear convex constraints to stabilize training and preserve generalization.
result FedCONST enhances feature transferability and robustness, achieving state-of-the-art performance.
Excessive reuse of holdout data can lead to overfitting. However, there is little concrete evidence of significant overfitting due to holdout reuse in popular multiclass benchmarks today. Known results show that, in the worst-case, revealing the accuracy of k adaptively chosen classifiers on a data set of size n al…
Excessive reuse of test data has become commonplace in today's machine learning workflows. Popular benchmarks, competitions, industrial scale tuning, among other applications, all involve test data reuse beyond guidance by statistical confidence bounds. Nonetheless, recent replication studies give evidence that popular…
Advanced Dropout improves DNN performance without requiring model-specific dropout techniques.
problem Overfitting in deep neural networks due to lack of data.
method Model-free, end-to-end training with adaptive dropout rate optimization.
result Advanced Dropout outperforms other techniques on various datasets.
IDS improves RLHF by smoothing reward data, enhancing model performance.
problem Reward model performance degrades and overoptimization hinders true objective.
method Iterative Data Smoothing (IDS) updates model and data labels during each epoch.
result IDS outperforms traditional methods in RLHF.
Study improves cancer classification using gene selection and projection methods.
problem Overfitting in high-dimensional microarray datasets for cancer classification.
method FSWOR technique, random projection, Kendall test, ensemble classifiers, LDA projection, Naïve Bayes.
result Achieved a test score of 96%, significantly outperforming existing methods.
Knowledge distillation improves model accuracy by mimicking teacher model probabilities.
problem Improving model accuracy through model compression.
method Casting knowledge distillation as a semiparametric inference problem, deriving new guarantees, and developing enhancements.
result Enhancements improve student performance by mitigating teacher overfitting and underfitting.
Smaller actor-critic models lead to performance degradation and overfitting, highlighting the critic's role in value underestimation.
problem Performance degradation and overfitting in actor-critic models with smaller actors.
method Broad empirical investigations and analyses of asymmetric actor-critic setups, exploring techniques to mitigate value underestimation.
result Value underestimation is a key cause of performance degradation in smaller actor-critic models, and the critic plays a crucial role in mitigating this.
Compressed LLM embeddings improve noisy regression tasks without overfitting.
problem Noisy regression tasks with high signal-to-noise ratios.
method Comparison of embedding compression techniques using autoencoder hidden representations.
result Compression improves performance on noisy tasks like financial return prediction.
Batch Normalization is a commonly used trick to improve the training of deep neural networks. These neural networks use L2 regularization, also called weight decay, ostensibly to prevent overfitting. However, we show that L2 regularization has no regularizing effect when combined with normalization. Instead, regulariza…
QuantBench benchmarks AI methods for quantitative investment.
problem Lack of a standardized benchmark for AI in quantitative investment.
method Developed an industrial-grade benchmark platform with standardization, flexibility, and full-pipeline coverage.
result Revealed critical research directions in AI for quantitative investment.
This paper reviews ML applications in finance, enhancing asset pricing models.
problem Limitations of traditional asset pricing models in complex market dynamics.
method Exploring ML models including supervised, unsupervised, semi-supervised, and reinforcement learning.
result Enhanced return prediction and portfolio optimization through ML integration.
New method uses generative models to improve phase retrieval stability.
problem Improving stability of solutions in phase retrieval problems.
method Unified reconstruction approach using generative models to mitigate overfitting.
result Mitigates overfitting to generative model for varying noise levels.
Despite recent success of deep network-based Reinforcement Learning (RL), it remains elusive to achieve human-level efficiency in learning novel tasks. While previous efforts attempt to address this challenge using meta-learning strategies, they typically suffer from sampling inefficiency with on-policy RL algorithms o…
New DP training ensures models behave similarly at training and test time.
problem Standard SGD training leads to inconsistent model behavior at training and test time.
method Differentially-Private (DP) training ensures WYSIWYG property through distributional generalization.
result DP training guarantees high-level WYSIWYG property, improving model robustness and privacy.
In this paper, we propose ℓp-norm regularized models to seek near-optimal sparse portfolios. These sparse solutions reduce the complexity of portfolio implementation and management. Theoretical results are established to guarantee the sparsity of the second-order KKT points of the ℓp-norm regularized models…
This paper investigates how synthetic data and overparameterization improve VAE generalization.
problem Improving generalization performance of Variational Autoencoders (VAEs).
method Investigates the effectiveness of synthetic data and overparameterization in VAEs.
result Training on synthetic data and using more parameters improves VAE generalization, inference, and robustness.
Improves domain adaptation by aligning source and target distributions and mitigating noisy labels.
problem Improving performance on target images with different acquisition conditions.
method Combines optimal transport, MixUp regularization, and robust loss for noisy labels.
result Improves domain adaptation performance on various benchmarks and real-world problems.
Paper proposes a new regularization method to prevent model degradation under distribution shifts.
problem Model performance degrades under distribution shifts.
method Supervised contrastive learning with heterogeneous similarity.
result The proposed method outperforms existing regularization methods on benchmark datasets.
DoubleEnsemble improves financial predictions by selecting key features and reweighting samples.
problem Overfitting and instability in financial data analysis.
method Sample reweighting and feature selection using learning trajectory and shuffling.
result DoubleEnsemble outperforms baseline methods in financial prediction tasks.
Spectral decoupling improves neural network generalization in medical imaging.
problem Poor generalization of neural networks trained on medical imaging data.
method Spectral decoupling, a regularization technique that encourages learning more features.
result Spectral decoupling increases network robustness and performance on external datasets.
Recently deep neural networks have shown their capacity to memorize training data, even with noisy labels, which hurts generalization performance. To mitigate this issue, we provide a simple but effective baseline method that is robust to noisy labels, even with severe noise. Our objective involves a variance regulariz…
InfoSFT improves LLMs by focusing on informative, medium-confidence tokens.
problem Overfitting to unlikely samples and degradation of prior capabilities in SFT.
method InfoSFT uses a principled weighting scheme to concentrate learning signals on medium-confidence tokens.
result InfoSFT improves generalization and preserves pre-existing capabilities over vanilla SFT and likelihood-weighted baselines.
In this paper, we propose Rogue-Gym, a simple and classic style roguelike game built for evaluating generalization in reinforcement learning (RL). Combined with the recent progress of deep neural networks, RL has successfully trained human-level agents without human knowledge in many games such as those for Atari 2600.…
KL annealing helps VAEs avoid posterior collapse and overfitting.
problem Posterior collapse and overfitting in VAEs.
method Theoretical analysis of learning dynamics with KL annealing.
result Posterior collapse is inevitable when β exceeds a threshold. Pseudo-label selection affects semi-supervised learning performance.
problem Selection of pseudo-labeled data impacts semi-supervised learning's generalization performance.
method Embedding pseudo-label selection into decision theory, deriving a novel selection criterion based on posterior predictive.
result BPLS (Bayesian pseudo-label selection) outperforms traditional methods in overfitting-prone data.
Our method upweights easy samples to mitigate forgetting in fine-tuning.
problem Catastrophic forgetting in fine-tuning pre-trained models.
method Sample weighting based on pre-trained model's losses.
result Our method reduces forgetting by up to 0.8% on MetaMathQA while preserving more accuracy on pre-training datasets.