Training Complex Models with Multi-Task Weak Supervision

Alexander Ratner, Braden Hancock, Jared Dunnmon, Frederic Sala, Shreyash Pandey, Christopher Ré

2018-10-05Paraphrase Identification Sentiment Analysis Natural Language Inference Semantic Textual Similarity Matrix Completion

Paper PDF Code(official)

Abstract

Snorkel MeTaL: A framework for training models with multi-task weak supervision

Results

Task	Dataset	Metric	Value	Model
Natural Language Inference	MultiNLI	Matched	87.6	Snorkel MeTaL (ensemble)
Natural Language Inference	MultiNLI	Mismatched	87.2	Snorkel MeTaL (ensemble)
Semantic Textual Similarity	Quora Question Pairs	Accuracy	89.9	Snorkel MeTaL(ensemble)
Semantic Textual Similarity	Quora Question Pairs	F1	73.1	Snorkel MeTaL(ensemble)
Sentiment Analysis	SST-2 Binary classification	Accuracy	96.2	Snorkel MeTaL(ensemble)
Paraphrase Identification	Quora Question Pairs	Accuracy	89.9	Snorkel MeTaL(ensemble)
Paraphrase Identification	Quora Question Pairs	F1	73.1	Snorkel MeTaL(ensemble)

Related Papers

AdaptiSent: Context-Aware Adaptive Attention for Multimodal Aspect-Based Sentiment Analysis2025-07-17 SemCSE: Semantic Contrastive Sentence Embeddings Using LLM-Generated Summaries For Scientific Abstracts2025-07-17 AI Wizards at CheckThat! 2025: Enhancing Transformer-Based Embeddings with Sentiment for Subjectivity Detection in News Articles2025-07-15 DCR: Quantifying Data Contamination in LLMs Evaluation2025-07-15 LRCTI: A Large Language Model-Based Framework for Multi-Step Evidence Retrieval and Reasoning in Cyber Threat Intelligence Credibility Verification2025-07-15 SentiDrop: A Multi Modal Machine Learning model for Predicting Dropout in Distance Learning2025-07-14 GNN-CNN: An Efficient Hybrid Model of Convolutional and Graph Neural Networks for Text Representation2025-07-10 DS@GT at CheckThat! 2025: Evaluating Context and Tokenization Strategies for Numerical Fact Verification2025-07-08