New method reveals how training data influence diffusion model outputs.
problem Difficulty in assessing training data impact on diffusion model outputs.
method Use of ensembles trained on carefully engineered splits of training data to identify influential training examples.
result Demonstrated the viability of ensembles as generative models and validity of assessing influence.
Regurgitative training with synthetic data harms LLM performance.
problem The impact of training LLMs with synthetic data generated by other LLMs.
method Fine-tuning GPT-3.5 with synthetic and real data in a machine translation task.
result Regurgitative training significantly reduces LLM performance.
In supervised machine learning for author name disambiguation, negative training data are often dominantly larger than positive training data. This paper examines how the ratios of negative to positive training data can affect the performance of machine learning algorithms to disambiguate author names in bibliographic …
Paper tackles model vulnerabilities by reconstructing training data.
problem Reconstructing training data from model parameters poses a security risk.
method Developed a mathematical framework and score matching method for both Bayesian and non-Bayesian models.
result First score matching framework for reconstructing data in Bayesian models.
Self-training outperforms pre-training on COCO object detection and segmentation datasets.
problem The effectiveness of pre-training in improving object detection and segmentation models is limited.
method Investigated self-training as an alternative method to utilize additional data.
result Self-training consistently improves model performance across various dataset sizes and data augmentation levels.
Fine-tuning with pre-training data improves performance.
problem Limited training data for tasks.
method Theoretical analysis of excess risk bound and selection of pre-training data subset.
result Improvement in generalization performance with pre-training data.
Captures data influence changes during training.
problem Traditional influence functions fail for modern training methods.
method Formalized trajectory-specific LOO influence, using data value embedding.
result Data influence varies by training stage, early and late stages have greater impact.
New algorithm efficiently trains machine learning models to atomic forces data.
problem Efficiently training machine learning models to large amounts of force data.
method Developed an efficient algorithm for training machine learning models to all available force data.
result Training to all available force data is only a few times more expensive than training to energies alone.
New MIP methods improve training of integer-valued neural networks.
problem Training integer-valued neural networks with limited data and resources.
method Formulated new MIP models to optimize training efficiency and handle more data.
result Significantly outperforms previous state-of-the-art methods in accuracy, training time, and data usage.
Proposes a strategy to train models with minimal labeled data.
problem Scarce and expensive labeled data for medical tasks.
method Recursive training strategy to use image-level annotations for pixel-level segmentation.
result Improved segmentation of intracranial hemorrhage in CT scans.
New method reconstructs significant parts of training data from neural networks.
problem Understanding and reconstructing training data from neural networks.
method Proposes a novel reconstruction scheme based on recent theoretical results about neural network training.
result Shows that a significant fraction of training data can be reconstructed from neural network parameters.
Accelerates data loading in deep neural network training by 30x.
problem Data loading is a bottleneck in deep neural network training.
method Locality-aware data loading method using software caches.
result More than 30x speedup in data loading.
DeepMimic trains neural networks using mostly unlabeled data.
problem Insufficient labeled data for deep learning.
method Mentor-Student approach with unlabeled data.
result Student model achieves mentor's performance without labeled data.
FR-Train improves fair and robust AI training by detecting and reducing poisoned data.
problem Training AI models that are fair and robust in the presence of data bias and poisoning.
method Mutual information-based adversarial training with an additional discriminator.
result FR-Train maintains fairness and accuracy even in the presence of poisoned data.
Deep neural networks can memorize training data even with just a few more parameters than samples.
problem Deep neural networks memorizing training data in mildly overparametrized regimes.
method Training neural networks with a number of parameters just a constant factor more than training samples.
result Neural networks can achieve 100% accuracy on training data in mildly overparametrized regimes.
Paper reconstructs training data from a single gradient query.
problem Privacy threats in federated learning due to model gradients.
method Provable attack using tensor decomposition.
result Training samples can be fully reconstructed from a single gradient query.
Self-training with noisy student-teacher boosts keyword spotting accuracy.
problem Robust keyword spotting in challenging conditions.
method Aggressive data augmentation and self-training with noisy student-teacher approach.
result Significant accuracy improvement in difficult conditions, up to 60%.
Extract synthetic data from pretrained models for tasks without training data.
problem Lack of training data for tasks requiring model initialization.
method Extract 'Data Impressions' from pretrained deep models' parameters.
result Data Impressions enable various tasks like unsupervised domain adaptation and continual learning.
Reweighting training data to better represent new tasks.
problem Deploying machine learning models to new tasks is challenging due to training data distribution.
method Formulate an exponential tilt distribution shift model and learn train data importance weights to minimize KL divergence.
result The learned train data weights improve target performance evaluation, fine-tuning, and model selection.
Study the effects of data parallelism and sparsity on neural network training.
problem Understanding the effects of data parallelism and sparsity on neural network training.
method Conducted extensive experiments and developed a theoretical analysis.
result Found a general scaling trend between batch size and number of training steps to convergence for the effect of data parallelism, and difficulty of training under sparsity.
Study evaluates how limited training data affects streamflow predictions.
problem Limited historical meteorological and streamflow data affects streamflow prediction accuracy.
method Evaluated tree- and LSTM-based models on CAMELS dataset with varying training data sizes and time spans.
result Tree- and LSTM-based models provide similarly accurate predictions on small datasets, but LSTMs are superior with more training data.
Recent years have witnessed amazing outcomes from "Big Models" trained by "Big Data". Most popular algorithms for model training are iterative. Due to the surging volumes of data, we can usually afford to process only a fraction of the training data in each iteration. Typically, the data are either uniformly sampled or…
Paper shows adversarial training can be fooled by new type of noise.
problem Adversarial training can be fooled by new types of noise.
method Designing ADVIN, a new type of inducing noise.
result ADVIN can degrade adversarial training robustness by 99.9%.
Support Vector Data Description (SVDD) is a popular outlier detection technique which constructs a flexible description of the input data. SVDD computation time is high for large training datasets which limits its use in big-data process-monitoring applications. We propose a new iterative sampling-based method for SVDD…
Measuring domain relevance of data and identifying or selecting well-fit domain data for machine translation (MT) is a well-studied topic, but denoising is not yet. Denoising is concerned with a different type of data quality and tries to reduce the negative impact of data noise on MT training, in particular, neural MT…
Sentiment analysis is a task that may suffer from a lack of data in certain cases, as the datasets are often generated and annotated by humans. In cases where data is inadequate for training discriminative models, generate models may aid training via data augmentation. Generative Adversarial Networks (GANs) are one suc…
Synthetic reference strings are as effective as real ones for training citation parsing models.
problem Lack of training data for citation parsing, especially with deep neural networks.
method Trained Grobid with human-labelled and synthetically created reference strings, and evaluated retraining and out-of-sample data impact.
result Synthetic and real reference strings are equally effective for training Grobid, with retraining improving performance.
Detects training data usage with radioactive data technique.
problem Detecting if a dataset was used to train a model.
method Radioactive data technique that makes imperceptible changes to detect training data usage.
result Can detect training data usage with high confidence (p<10^-4) even with small amounts of radioactive data.
In this paper we propose the use of Generative Adversarial Networks (GAN) to generate artificial training data for machine learning tasks. The generation of artificial training data can be extremely useful in situations such as imbalanced data sets, performing a role similar to SMOTE or ADASYN. It is also useful when t…
Develops a robust training framework to detect backdoor attacks in DNNs.
problem Vulnerability of DNNs to backdoor attacks by poisoned training data.
method Collider framework selects prominent samples based on geometric structures and coreset selection objective.
result Significantly reduces backdoor success rate in various poisoned datasets.
ES improves training efficiency by dynamically selecting data samples.
problem Efficiently selecting informative data samples for faster learning.
method Evolved Sampling (ES) dynamically selects data samples based on loss dynamics and differences.
result ES achieves significant training acceleration without compromising model performance.
A novel approach of training data augmentation and domain adaptation is presented to support machine learning applications for cognitive radio. Machine learning provides effective tools to automate cognitive radio functionalities by reliably extracting and learning intrinsic spectrum dynamics. However, there are two im…
Datamodels predict model outcomes from training data subsets.
problem Understanding model behavior from training data.
method Conceptual framework for analyzing model behavior using datamodels.
result Simple linear datamodels can accurately predict model outputs.
The paper studies stability of generative models trained on mixed data.
problem Training generative models on mixed datasets (real and synthetic data).
method Developed a framework to rigorously study stability under specific conditions.
result Proved the stability of iterative training under certain conditions.
Framework for few-shot relation classification with minimal training data.
problem Few-shot relation classification with limited training data.
method Meta-learning framework that combines instance and support knowledge.
result Framework outperforms state-of-the-art results and achieves competitive performance with large training data.
Denoising autoencoders (DAEs) are powerful deep learning models used for feature extraction, data generation and network pre-training. DAEs consist of an encoder and decoder which may be trained simultaneously to minimise a loss (function) between an input and the reconstruction of a corrupted version of the input. The…
This paper proposes a new D2D data sharing approach to improve distributed machine learning training speed.
problem Straggler dilemma in distributed edge learning.
method Proposes a D2D data sharing approach to balance computation loads and optimize radio resource allocation.
result Significantly reduces training delay and enhances training accuracy in non-i.i.d. data environments.
Different types of training data have led to numerous schemes for supervised classification. Current learning techniques are tailored to one specific scheme and cannot handle general ensembles of training data. This paper presents a unifying framework for supervised classification with general ensembles of training dat…
Doubly robust self-training improves semi-supervised learning by balancing labeled and pseudo-labeled data.
problem Improving semi-supervised learning performance with limited labeled data.
method Introduces doubly robust self-training, a method that combines labeled and pseudo-labeled data to balance between labeled-only and pseudo-labeled-only training.
result Demonstrates superior performance of doubly robust self-training on ImageNet and nuScenes datasets.
PCRs compress data for deep learning, reducing training time.
problem Efficiently training deep learning models over large datasets.
method Combining progressive compression with an efficient storage layout.
result PCRs can tolerate up to 50% compression without significantly affecting training accuracy.
The paper analyzes how generated data improves adversarial training in high-dimensional regression.
problem Improving adversarial training in high-dimensional regression.
method Theoretical analysis of a two-stage training approach with generated data and pseudo-labels.
result Two-stage adversarial training achieves better performance than ridgeless training in high-dimensional linear regression.
Stabilizes GAN training with limited data.
problem Overfitting in GANs with scarce data.
method Adaptive discriminator augmentation.
result Good results possible with few thousand images.
Paper tackles model collapse in recursive generative models using a weighted training scheme.
problem Model collapse in recursive generative models trained on synthetic data.
method Iteratively trains models on real and synthetic data, evaluates weighted training schemes.
result Optimal weighting scheme for synthetic data follows a unified expression, revealing a trade-off with model performance.
Adversarial training leads to clean data generalization with significant robust overfitting gap.
problem Significant robust generalization gap in adversarial training.
method Two theoretical views: representation complexity and training dynamics.
result ReLU nets with O(ND) extra parameters can achieve CGRO. The paper develops an algorithm to select a subset of training data for efficient regression models.
problem Designing an efficient algorithm for selecting a subset of training data to train regression models quickly without sacrificing accuracy.
method The paper tackles this problem by formulating it as a minimization of training loss with respect to both trainable parameters and subset of training data, subject to error bounds on the validation set. They use a novel problem formulation and represent it with simplified constraints using the dual of the original training problem. They then develop SELCON, an efficient majorization-minimization algorithm for data subset selection, which admits an approximation guarantee.
result The experiments show that SELCON trades off accuracy and efficiency more effectively than the current state-of-the-art.
Knowledge distillation deals with the problem of training a smaller model (Student) from a high capacity source model (Teacher) so as to retain most of its performance. Existing approaches use either the training data or meta-data extracted from it in order to train the Student. However, accessing the dataset on which …
Unified method for multi-defect microscopy image restoration with limited training data.
problem Challenges in applying deep learning methods due to limited training data for multi-defect microscopy images.
method Two-stage approach: data augmentation with GAN and conditional GAN training.
result Proposed method gives comparable or superior results to existing methods in image quality restoration.
Supervised learning from training data with imbalanced class sizes, a commonly encountered scenario in real applications such as anomaly/fraud detection, has long been considered a significant challenge in machine learning. Motivated by recent progress in curriculum and self-paced learning, we propose to adopt a semi-s…