Search Results for author: Alan Fern

Found 40 papers, 12 papers with code

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

no code implementations4 Jun 2022 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.

Natural Language Processing

Offline Policy Comparison with Confidence: Benchmarks and Baselines

1 code implementation22 May 2022 Anurag Koul, Mariano Phielipp, Alan Fern

Decision makers often wish to use offline historical data to compare sequential-action policies at various world states.

Sim-to-Real Learning for Bipedal Locomotion Under Unsensed Dynamic Loads

no code implementations9 Apr 2022 Jeremy Dao, Kevin Green, Helei Duan, Alan Fern, Jonathan Hurst

We show that prior RL policies trained for unloaded locomotion fail for some loads and that simply training in the context of loads is enough to result in successful and improved policies.

Identifying Reasoning Flaws in Planning-Based RL Using Tree Explanations

no code implementations28 Sep 2021 Kin-Ho Lam, Zhengxian Lin, Jed Irvine, Jonathan Dodge, Zeyad T Shureih, Roli Khanna, Minsuk Kahng, Alan Fern

We describe a user interface and case study, where a small group of AI experts and developers attempt to identify reasoning flaws due to inaccurate agent learning.

Decision Making

From Heatmaps to Structural Explanations of Image Classifiers

no code implementations13 Sep 2021 Li Fuxin, Zhongang Qi, Saeed Khorram, Vivswan Shitole, Prasad Tadepalli, Minsuk Kahng, Alan Fern

This paper summarizes our endeavors in the past few years in terms of explaining image classifiers, with the aim of including negative results and insights we have gained.

Out-of-Distribution Dynamics Detection: RL-Relevant Benchmarks and Results

2 code implementations11 Jul 2021 Mohamad H Danesh, Alan Fern

This is relevant to applications in control, reinforcement learning (RL), and multi-variate time-series, where changes to test time dynamics can impact the performance of learning controllers/predictors in unknown ways.

Time Series

Piecewise-constant Neural ODEs

1 code implementation11 Jun 2021 Sam Greydanus, Stefan Lee, Alan Fern

Neural networks are a popular tool for modeling sequential data but they generally do not treat time as a continuous variable.

Deep Convolution for Irregularly Sampled Temporal Point Clouds

no code implementations1 May 2021 Erich Merrill, Stefan Lee, Li Fuxin, Thomas G. Dietterich, Alan Fern

We consider the problem of modeling the dynamics of continuous spatial-temporal processes represented by irregular samples through both space and time.

Starcraft Starcraft II

Jumpy Recurrent Neural Networks

no code implementations1 Jan 2021 Samuel James Greydanus, Stefan Lee, Alan Fern

This structure enables our model to jump over long time intervals while retaining the ability to produce fine-grained or continuous-time predictions when necessary.

Time Series

Optimizing Discrete Spaces via Expensive Evaluations: A Learning to Search Framework

no code implementations14 Dec 2020 Aryan Deshwal, Syrine Belakaria, Janardhan Rao Doppa, Alan Fern

We consider the problem of optimizing expensive black-box functions over discrete spaces (e. g., sets, sequences, graphs).

Sim-to-Real Learning of All Common Bipedal Gaits via Periodic Reward Composition

no code implementations2 Nov 2020 Jonah Siekmann, Yesh Godse, Alan Fern, Jonathan Hurst

We study the problem of realizing the full spectrum of bipedal locomotion on a real robot with sim-to-real reinforcement learning (RL).


Dream and Search to Control: Latent Space Planning for Continuous Control

1 code implementation19 Oct 2020 Anurag Koul, Varun V. Kumar, Alan Fern, Somdeb Majumdar

Learning and planning with latent space dynamics has been shown to be useful for sample efficiency in model-based reinforcement learning (MBRL) for discrete and continuous control tasks.

Continuous Control Model-based Reinforcement Learning

DeepAveragers: Offline Reinforcement Learning by Solving Derived Non-Parametric MDPs

2 code implementations ICLR 2021 Aayam Shrestha, Stefan Lee, Prasad Tadepalli, Alan Fern

We study an approach to offline reinforcement learning (RL) based on optimally solving finitely-represented MDPs derived from a static dataset of experience.

Offline RL reinforcement-learning

Contrastive Explanations for Reinforcement Learning via Embedded Self Predictions

1 code implementation ICLR 2021 Zhengxian Lin, Kim-Ho Lam, Alan Fern

We investigate a deep reinforcement learning (RL) architecture that supports explaining why a learned agent prefers one action over another.


Re-understanding Finite-State Representations of Recurrent Policy Networks

1 code implementation6 Jun 2020 Mohamad H. Danesh, Anurag Koul, Alan Fern, Saeed Khorram

We introduce an approach for understanding control policies represented as recurrent neural networks.

Atari Games

The Choice Function Framework for Online Policy Improvement

no code implementations1 Oct 2019 Murugeswari Issakkimuthu, Alan Fern, Prasad Tadepalli

There are notable examples of online search improving over hand-coded or learned policies (e. g. AlphaZero) for sequential decision making.

Decision Making

Sample-Based Point Cloud Decoder Networks

no code implementations25 Sep 2019 Erich Merrill, Alan Fern

In this work, we investigate decoder architectures that more closely match the semantics of variable sized point clouds.

Explaining Reinforcement Learning to Mere Mortals: An Empirical Study

no code implementations22 Mar 2019 Andrew Anderson, Jonathan Dodge, Amrita Sadarangani, Zoe Juozapaitis, Evan Newman, Jed Irvine, Souti Chattopadhyay, Alan Fern, Margaret Burnett

We present a user study to investigate the impact of explanations on non-experts' understanding of reinforcement learning (RL) agents.


Interactive Naming for Explaining Deep Neural Networks: A Formative Study

no code implementations18 Dec 2018 Mandana Hamidi-Haines, Zhongang Qi, Alan Fern, Fuxin Li, Prasad Tadepalli

For this purpose, we developed a user interface for "interactive naming," which allows a human annotator to manually cluster significant activation maps in a test set into meaningful groups called "visual concepts".

General Classification

Learning Finite State Representations of Recurrent Policy Networks

no code implementations ICLR 2019 Anurag Koul, Sam Greydanus, Alan Fern

Recurrent neural networks (RNNs) are an effective representation of control policies for a wide range of reinforcement and imitation learning problems.

Atari Games Imitation Learning

Open Category Detection with PAC Guarantees

1 code implementation ICML 2018 Si Liu, Risheek Garrepalli, Thomas G. Dietterich, Alan Fern, Dan Hendrycks

Further, while there are algorithms for open category detection, there are few empirical results that directly report alien detection rates.

Visualizing and Understanding Atari Agents

3 code implementations ICML 2018 Sam Greydanus, Anurag Koul, Jonathan Dodge, Alan Fern

While deep reinforcement learning (deep RL) agents are effective at maximizing rewards, it is often unclear what strategies they use to do so.


Incorporating Feedback into Tree-based Anomaly Detection

2 code implementations30 Aug 2017 Shubhomoy Das, Weng-Keen Wong, Alan Fern, Thomas G. Dietterich, Md Amran Siddiqui

Unfortunately, in realworld applications, this process can be exceedingly difficult for the analyst since a large fraction of high-ranking anomalies are false positives and not interesting from the application perspective.

Anomaly Detection

Budget-Aware Deep Semantic Video Segmentation

no code implementations CVPR 2017 Behrooz Mahasseni, Sinisa Todorovic, Alan Fern

In this work, we study a poorly understood trade-off between accuracy and runtime costs for deep semantic video segmentation.

Action Detection Activity Detection +3

Approximate Policy Iteration for Budgeted Semantic Video Segmentation

no code implementations26 Jul 2016 Behrooz Mahasseni, Sinisa Todorovic, Alan Fern

Our second contribution is the algorithm for learning a policy for the sparse selection of supervoxels and their descriptors for budgeted CRF inference.

Video Segmentation Video Semantic Segmentation

Schema Independent Relational Learning

no code implementations16 Aug 2015 Jose Picado, Arash Termehchy, Alan Fern, Parisa Ataei

In this paper, we introduce and formalize the property of schema independence of relational learning algorithms, and study both the theoretical and empirical dependence of existing algorithms on the common class of (de) composition schema transformations.

Relational Reasoning

Person Count Localization in Videos From Noisy Foreground and Detections

no code implementations CVPR 2015 Sheng Chen, Alan Fern, Sinisa Todorovic

This problem is a middle-ground between frame-level person counting, which does not localize counts, and person detection aimed at perfectly localizing people with count-one detections.

Human Detection Video Understanding

A Meta-Analysis of the Anomaly Detection Problem

1 code implementation3 Mar 2015 Andrew Emmott, Shubhomoy Das, Thomas Dietterich, Alan Fern, Weng-Keen Wong

The intended contributions of this article are many; in addition to providing a large publicly-available corpus of anomaly detection benchmarks, we provide an ontology for describing anomaly detection contexts, a methodology for controlling various aspects of benchmark creation, guidelines for future experimental design and a discussion of the many potential pitfalls of trying to measure success in this field.

Anomaly Detection Experimental Design +1

Sequential Feature Explanations for Anomaly Detection

no code implementations28 Feb 2015 Md Amran Siddiqui, Alan Fern, Thomas G. Dietterich, Weng-Keen Wong

An SFE of an anomaly is a sequence of features, which are presented to the analyst one at a time (in order) until the information contained in the highlighted features is enough for the analyst to make a confident judgement about the anomaly.

Anomaly Detection

Multi-Object Tracking via Constrained Sequential Labeling

no code implementations CVPR 2014 Sheng Chen, Alan Fern, Sinisa Todorovic

This paper presents a new approach to tracking people in crowded scenes, where people are subject to long-term (partial) occlusions and may assume varying postures and articulations.

Multi-Object Tracking

Coactive Learning for Locally Optimal Problem Solving

no code implementations18 Apr 2014 Robby Goetschalckx, Alan Fern, Prasad Tadepalli

Coactive learning is an online problem solving setting where the solutions provided by a solver are interactively improved by a domain expert, which in turn drives learning.

Symbolic Opportunistic Policy Iteration for Factored-Action MDPs

no code implementations NeurIPS 2013 Aswin Raghavan, Roni Khardon, Alan Fern, Prasad Tadepalli

We address the scalability of symbolic planning under uncertainty with factored states and actions.

Batch Active Learning via Coordinated Matching

no code implementations27 Jun 2012 Javad Azimi, Alan Fern, Xiaoli Zhang-Fern, Glencora Borradaile, Brent Heeringa

Most prior work on active learning of classifiers has focused on sequentially selecting one unlabeled example at a time to be labeled in order to reduce the overall labeling effort.

Active Learning Combinatorial Optimization

Budgeted Optimization with Concurrent Stochastic-Duration Experiments

no code implementations NeurIPS 2011 Javad Azimi, Alan Fern, Xiaoli Z. Fern

This paper defines a novel problem formulation with the following important extensions: 1) allowing for concurrent experiments; 2) allowing for stochastic experiment durations; and 3) placing constraints on both the total number of experiments and the total experimental time.

Autonomous Learning of Action Models for Planning

no code implementations NeurIPS 2011 Neville Mehta, Prasad Tadepalli, Alan Fern

This paper introduces two new frameworks for learning action models for planning.

Batch Bayesian Optimization via Simulation Matching

no code implementations NeurIPS 2010 Javad Azimi, Alan Fern, Xiaoli Z. Fern

Bayesian optimization methods are often used to optimize unknown functions that are costly to evaluate.

A Computational Decision Theory for Interactive Assistants

no code implementations NeurIPS 2010 Alan Fern, Prasad Tadepalli

A variation of this policy is shown to achieve worst-case regret that is logarithmic in the number of goals for any goal distribution.

Cannot find the paper you are looking for? You can Submit a new open access paper.