Author name cluster

Jed Irvine

Possible papers associated with this exact author name in Arrow. This page groups case-insensitive exact name matches and is not a full identity disambiguation profile.

4 papers

2 author rows

ICAPS Conference 2022 Conference Paper

Beyond Value: CheckList for Testing Inferences in Planning-Based RL

Kin-Ho Lam
Delyar Tabatabai
Jed Irvine
Donald Bertucci
Anita Ruangrotsakun
Minsuk Kahng
Alan Fern

Reinforcement learning (RL) agents are commonly evaluated via their expected value over a distribution of test scenarios. Unfortunately, this evaluation approach provides limited evidence for post-deployment generalization beyond the test distribution. In this paper, we address this limitation by extending the recent CheckList testing methodology from natural language processing to planning-based RL. Specifically, we consider testing RL agents that make decisions via online tree search using a learned transition model and value function. The key idea is to improve the assessment of future performance via a CheckList approach for exploring and assessing the agent's inferences during tree search. The approach provides the user with an interface and general query-rule mechanism for identifying potential inference flaws and validating expected inference invariances. We present a user study involving knowledgeable AI researchers using the approach to evaluate an agent trained to play a complex real-time strategy game. The results show the approach is effective in allowing users to identify previously-unknown flaws in the agent's reasoning. In addition, our analysis provides insight into how AI experts use this type of testing approach, which may help improve future instantiations.

Details

IJCAI Conference 2019 Conference Paper

Explaining Reinforcement Learning to Mere Mortals: An Empirical Study

Andrew Anderson
Jonathan Dodge
Amrita Sadarangani
Zoe Juozapaitis
Evan Newman
Jed Irvine
Souti Chattopadhyay
Alan Fern

We present a user study to investigate the impact of explanations on non-experts? understanding of reinforcement learning (RL) agents. We investigate both a common RL visualization, saliency maps (the focus of attention), and a more recent explanation type, reward-decomposition bars (predictions of future types of rewards). We designed a 124 participant, four-treatment experiment to compare participants? mental models of an RL agent in a simple Real-Time Strategy (RTS) game. Our results show that the combination of both saliency and reward bars were needed to achieve a statistically significant improvement in mental model score over the control. In addition, our qualitative analysis of the data reveals a number of effects for further study.

PDF Details

AAAI Conference 2013 Conference Paper

Approximate Bayesian Inference for Reconstructing Velocities of Migrating Birds from Weather Radar

Daniel Sheldon
Andrew Farnsworth
Jed Irvine
Benjamin Van Doren
Kevin Webb
Thomas Dietterich
Steve Kelling

Archived data from the WSR-88D network of weather radars in the US hold detailed information about the continent-scale migratory movements of birds over the last 20 years. However, significant technical challenges must be overcome to understand this information and harness its potential for science and conservation. We present an approximate Bayesian inference algorithm to reconstruct the velocity fields of birds migrating in the vicinity of a radar station. This is part of a larger project to quantify bird migration at large scales using weather radar data.

PDF Details

AAAI Conference 2006 Conference Paper

Predicting Task-Specific Webpages for Revisiting

Arwen Twinkle Lettkeman
Jed Irvine

With the increased use of the web has come a corresponding increase in information overload that users face when trying to locate specific webpages, especially as a majority of visits to webpages are revisits. While automatically created browsing history lists offer a potential low-cost solution to re-locating webpages, even short browsing sessions generate a glut of webpages that do not relate to the user's information need or have no revisit value. We address how we can better support web users who want to return to information on a webpage that they have previously visited by building more useful history lists. The paper reports on a combination technique that semi-automatically segments the webpage browsing history list into tasks, applies heuristics to remove webpages that carry no intrinsic revisit value, and uses a learning model, sensitive to individual users and tasks, that predicts which webpages are likely to be revisited again. We present results from an empirical evaluation that report the likely revisit need of users and that show that adequate overall prediction accuracy can be achieved. This approach can be used to increase utility of history lists by removing information overload to users when revisiting webpages.

PDF Details