Autonomous Curiosity for Real-Time Training Onboard Robotic Agents

08/29/2021
by   Ervin Teng, et al.
0

Learning requires both study and curiosity. A good learner is not only good at extracting information from the data given to it, but also skilled at finding the right new information to learn from. This is especially true when a human operator is required to provide the ground truth - such a source should only be queried sparingly. In this work, we address the problem of curiosity as it relates to online, real-time, human-in-the-loop training of an object detection algorithm onboard a robotic platform, one where motion produces new views of the subject. We propose a deep reinforcement learning approach that decides when to ask the human user for ground truth, and when to move. Through a series of experiments, we demonstrate that our agent learns a movement and request policy that is at least 3x more effective at using human user interactions to train an object detector than untrained approaches, and is generalizable to a variety of subjects and environments.

READ FULL TEXT

page 6

page 8

research
02/05/2019

Learning to Learn in Simulation

Deep learning often requires the manual collection and annotation of a t...
research
10/07/2021

Improving Robot-Centric Learning from Demonstration via Personalized Embeddings

Learning from demonstration (LfD) techniques seek to enable novice users...
research
11/29/2017

Real-Time System for Human Activity Analysis

We propose a real-time human activity analysis system, where a user's ac...
research
04/15/2021

Human-in-the-Loop Deep Reinforcement Learning with Application to Autonomous Driving

Due to the limited smartness and abilities of machine intelligence, curr...
research
12/09/2020

Route Reconstruction from Traffic Flow via Representative Trajectories

Understanding human mobility is an important aspect of traffic analysis ...
research
03/27/2018

ClickBAIT-v2: Training an Object Detector in Real-Time

Modern deep convolutional neural networks (CNNs) for image classificatio...
research
06/21/2014

Interactively Test Driving an Object Detector: Estimating Performance on Unlabeled Data

In this paper, we study the problem of `test-driving' a detector, i.e. a...

Please sign up or login with your details

Forgot password? Click here to reset