Training Data

Training data is crucial for machine learning model development, with current research focusing on improving data quality, efficiency, and mitigating biases. Active areas include generating synthetic data to address scarcity or privacy concerns, developing algorithms to optimize data selection and usage (e.g., self-paced learning, active learning), and mitigating issues like data contamination and imbalance through techniques such as data augmentation, selective parameter merging, and novel loss functions. The quality and characteristics of training data significantly impact model performance, generalization, and robustness, influencing various applications from natural language processing and image recognition to scientific computing and medical diagnosis.

Papers

November 29, 2022

Every Node Counts: Improving the Training of Graph Neural Networks on Node Classification
Moshe Eliasof, Eldad Haber, Eran Treister
Graph Neural Network Training Data Node Classification Random Node Node Classification Datasets

November 26, 2022

The Impact of Racial Distribution in Training Data on Face Recognition Bias: A Closer Look
Manideep Kolla, Aravinth Savadamuthu
Training Data Global Impact Face Recognition Face Recognition Model Closer Look Facial Impression Bias

November 25, 2022

November 23, 2022

Improving dermatology classifiers across populations using images generated by large diffusion models
Luke W. Sagers, James A. Diao, Matthew Groh, Pranav Rajpurkar, Adewole S. Adamson, Arjun K. Manrai
Training Data Text to Image Diffusion Model Synthetic Image Digital Image Skin Cancer Photorealistic Image Na\"ive Population Large Diffusion Model

November 22, 2022

Noise-resilient approach for deep tomographic imaging
Zhen Guo, Zhiguang Liu, Qihang Zhang, George Barbastathis, Michael E. Glinsky
Training Data X Ray Robust Reconstruction Noise Resistant Property Neural Adaptive Tomography

November 20, 2022

Learning from Long-Tailed Noisy Data with Sample Selection and Balanced Loss
Lefan Zhang, Zhang-Hao Tian, Wujun Zhou, Wei Wang
LeArning Abstract Training Data Noisy Data Long Tailed Data Sample Selection Balanced Loss

November 19, 2022

Quantifying Human Bias and Knowledge to guide ML models during Training
Hrishikesh Viswanath, Andrey Shor, Yoshimasa Kitaguchi
Training Data Machine Learning Model Knowledge Based Imbalanced Data Crowdsourcing Context Length Bias Quantifying Bias Initial Weight

November 17, 2022

I Can't Believe There's No Images! Learning Visual Tasks Using only Language Supervision
Sophia Gu, Christopher Clark, Aniruddha Kembhavi
Training Data Visual Task Language Supervision Contrastive Model Visual Entailment News Caption

November 16, 2022

Creative divergent synthesis with generative models
Axel Chemla--Romeu-Santos, Philippe Esling
Generative Model Training Data Adversarial Learning Data Distribution Training Objective Generation Capability Divergent Thinking

November 15, 2022

November 14, 2022

SNIPER Training: Single-Shot Sparse Training for Text-to-Speech
Perry Lam, Huayun Zhang, Nancy F. Chen, Berrak Sisman, Dorien Herremans
Training Data Text to Speech Speech to Text Sparse Training High Sparsity Dense Training Sparsity Aware

November 13, 2022

Differentially Private Vertical Federated Learning
Thilina Ranbaduge, Ming Ding
Deep Learning Training Data Differential Privacy Vertical Federated Learning Machine Learning Component

November 12, 2022

Multi-Speaker and Wide-Band Simulated Conversations as Training Data for End-to-End Neural Diarization
Federico Landini, Mireia Diez, Alicia Lozano-Diez, Lukáš Burget
Training Data Voice Activity Detection End to End Model Diarization System Refined Diarization End to End Neural Diarization

November 10, 2022

Training Data

Papers

Every Node Counts: Improving the Training of Graph Neural Networks on Node Classification

The Impact of Racial Distribution in Training Data on Face Recognition Bias: A Closer Look

MS-PS: A Multi-Scale Network for Photometric Stereo With a New Comprehensive Training Dataset

Training Data Improvement for Image Forgery Detection using Comprint

A Survey of Learning Curves with Bad Behavior: or How More Data Need Not Lead to Better Performance

Combating noisy labels in object detection datasets

TrustGAN: Training safe and trustworthy deep learning models through generative adversarial networks

Improving dermatology classifiers across populations using images generated by large diffusion models

Noise-resilient approach for deep tomographic imaging

Learning from Long-Tailed Noisy Data with Sample Selection and Balanced Loss

Quantifying Human Bias and Knowledge to guide ML models during Training

I Can't Believe There's No Images! Learning Visual Tasks Using only Language Supervision

Creative divergent synthesis with generative models

Dynamic-Pix2Pix: Noise Injected cGAN for Modeling Input and Target Domain Joint Distributions with Limited Training Data

Will Large-scale Generative Models Corrupt Future Datasets?

SNIPER Training: Single-Shot Sparse Training for Text-to-Speech

Differentially Private Vertical Federated Learning

Multi-Speaker and Wide-Band Simulated Conversations as Training Data for End-to-End Neural Diarization

The CRINGE Loss: Learning what language not to model

A Comprehensive Survey on Distributed Training of Graph Neural Networks