DeepAI AI Chat
Log In Sign Up

Cross-task weakly supervised learning from instructional videos

by   Dimitri Zhukov, et al.

In this paper we investigate learning visual models for the steps of ordinary tasks using weak supervision via instructional narrations and an ordered list of steps instead of strong supervision via temporal annotations. At the heart of our approach is the observation that weakly supervised learning may be easier if a model shares components while learning different steps: `pour egg' should be trained jointly with other tasks involving `pour' and `egg'. We formalize this in a component model for recognizing steps and a weakly supervised learning framework that can learn this model under temporal constraints from narration and the list of steps. Past data does not permit systematic studying of sharing and so we also gather a new dataset, CrossTask, aimed at assessing cross-task sharing. Our experiments demonstrate that sharing across tasks improves performance, especially when done at the component level and that our component model can parse previously unseen tasks by virtue of its compositionality.


page 1

page 6

page 7

page 8


Bandit Label Inference for Weakly Supervised Learning

The scarcity of data annotated at the desired level of granularity is a ...

Weakly-supervised Action Localization with Background Modeling

We describe a latent approach that learns to detect actions in long sequ...

From Weakly Supervised Learning to Biquality Learning, a brief introduction

The field of Weakly Supervised Learning (WSL) has recently seen a surge ...

Unified Risk Analysis for Weakly Supervised Learning

Among the flourishing research of weakly supervised learning (WSL), we r...

Weakly Supervised Learning Meets Ride-Sharing User Experience Enhancement

Weakly supervised learning aims at coping with scarce labeled data. Prev...

Constrained Deep Weak Supervision for Histopathology Image Segmentation

In this paper, we develop a new weakly-supervised learning algorithm to ...

From Weakly Supervised Learning to Active Learning

Applied mathematics and machine computations have raised a lot of hope s...