Webly Supervised Concept Expansion for General Purpose Vision Models

Amita Kamath, Christopher Clark, Tanmay Gupta, Eric Kolve, Derek Hoiem, Aniruddha Kembhavi

Research output: Chapter in Book/Report/Conference proceedingConference contribution


General Purpose Vision (GPV) systems are models that are designed to solve a wide array of visual tasks without requiring architectural changes. Today, GPVs primarily learn both skills and concepts from large fully supervised datasets. Scaling GPVs to tens of thousands of concepts by acquiring data to learn each concept for every skill quickly becomes prohibitive. This work presents an effective and inexpensive alternative: learn skills from supervised datasets, learn concepts from web image search, and leverage a key characteristic of GPVs: the ability to transfer visual knowledge across skills. We use a dataset of 1M+ images spanning 10k+ visual concepts to demonstrate webly-supervised concept expansion for two existing GPVs (GPV-1 and VL-T5) on 3 benchmarks: 5 Coco-based datasets (80 primary concepts), a newly curated series of 5 datasets based on the OpenImages and VisualGenome repositories (∼ 500 concepts), and the Web-derived dataset (10k+ concepts). We also propose a new architecture, GPV-2 that supports a variety of tasks — from vision tasks like classification and localization to vision+language tasks like QA and captioning, to more niche ones like human-object interaction detection. GPV-2 benefits hugely from web data and outperforms GPV-1 and VL-T5 across these benchmarks. Our data, code, and web demo are available at https://prior.allenai.org/projects/gpv2.

Original languageEnglish (US)
Title of host publicationComputer Vision – ECCV 2022 - 17th European Conference, 2022, Proceedings
EditorsShai Avidan, Gabriel Brostow, Moustapha Cissé, Giovanni Maria Farinella, Tal Hassner
Number of pages20
ISBN (Print)9783031200588
StatePublished - 2022
Event17th European Conference on Computer Vision, ECCV 2022 - Tel Aviv, Israel
Duration: Oct 23 2022Oct 27 2022

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume13696 LNCS
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349


Conference17th European Conference on Computer Vision, ECCV 2022
CityTel Aviv


  • General purpose vision systems
  • Webly supervised data

ASJC Scopus subject areas

  • Theoretical Computer Science
  • General Computer Science


Dive into the research topics of 'Webly Supervised Concept Expansion for General Purpose Vision Models'. Together they form a unique fingerprint.

Cite this