Learning from Human Feedback: Challenges for Real-World Reinforcement Learning in NLP

11/04/2020 ∙ by Julia Kreutzer, et al. ∙ 73 ∙ share

Large volumes of interaction logs can be collected from NLP systems that are deployed in the real world. How can this wealth of information be leveraged? Using such interaction logs in an offline reinforcement learning (RL) setting is a promising approach. However, due to the nature of NLP tasks and the constraints of production systems, a series of challenges arise. We present a concise overview of these challenges and discuss possible solutions.

READ FULL TEXT

Challenges of Real-World Reinforcement Learning

Reinforcement learning (RL) has proven its worth in a series of artifici...
04/29/2019 ∙ by Gabriel Dulac-Arnold, et al. ∙ 0 ∙ share

read it
Using Conformity to Probe Interaction Challenges in XR Collaboration

The concept of a conformity spectrum is introduced to describe the degre...
04/10/2020 ∙ by Jeremy Hartmann, et al. ∙ 0 ∙ share

read it
An empirical investigation of the challenges of real-world reinforcement learning

Reinforcement learning (RL) has proven its worth in a series of artifici...
03/24/2020 ∙ by Gabriel Dulac-Arnold, et al. ∙ 9 ∙ share

read it
Balancing Constraints and Rewards with Meta-Gradient D4PG

Deploying Reinforcement Learning (RL) agents to solve real-world applica...
10/13/2020 ∙ by Dan A. Calian, et al. ∙ 0 ∙ share

read it
Batch-Constrained Distributional Reinforcement Learning for Session-based Recommendation

Most of the existing deep reinforcement learning (RL) approaches for ses...
12/16/2020 ∙ by Diksha Garg, et al. ∙ 0 ∙ share

read it
Explaining Conditions for Reinforcement Learning Behaviors from Real and Imagined Data

The deployment of reinforcement learning (RL) in the real world comes wi...
11/17/2020 ∙ by Aastha Acharya, et al. ∙ 0 ∙ share

read it
Robust Constrained Reinforcement Learning for Continuous Control with Model Misspecification

Many real-world physical control systems are required to satisfy constra...
10/20/2020 ∙ by Daniel J. Mankowitz, et al. ∙ 0 ∙ share

read it

Learning from Human Feedback: Challenges for Real-World Reinforcement Learning in NLP

Authors

Challenges of Real-World Reinforcement Learning

Using Conformity to Probe Interaction Challenges in XR Collaboration

An empirical investigation of the challenges of real-world reinforcement learning

Balancing Constraints and Rewards with Meta-Gradient D4PG

Batch-Constrained Distributional Reinforcement Learning for Session-based Recommendation

Explaining Conditions for Reinforcement Learning Behaviors from Real and Imagined Data

Robust Constrained Reinforcement Learning for Continuous Control with Model Misspecification

This week in AI