Skip to main content

Randy Goebel

University of Alberta, Computing Science, Faculty Member

Followers

139

Following

35

Co-authors

28

Public Views

Address: Edmonton, Alberta, Canada

less

Interests

Uploads

Papers by Randy Goebel

NeurIPS 2022 Competition: Driving SMARTS

arXiv (Cornell University), Nov 14, 2022

Driving SMARTS is a regular competition designed to tackle problems caused by the distribution sh... more Driving SMARTS is a regular competition designed to tackle problems caused by the distribution shift in dynamic interaction contexts that are prevalent in realworld autonomous driving (AD). The proposed competition supports methodologically diverse solutions, such as reinforcement learning (RL) and offline learning methods, trained on a combination of naturalistic AD data and open-source simulation platform SMARTS. The two-track structure allows focusing on different aspects of the distribution shift. Track 1 is open to any method and will give ML researchers with different backgrounds an opportunity to solve a real-world autonomous driving challenge. Track 2 is designed for strictly offline learning methods. Therefore, direct comparisons can be made between different methods with the aim to identify new promising research directions. The proposed setup consists of 1) realistic traffic generated using real-world data and micro simulators to ensure fidelity of the scenarios, 2) framework accommodating diverse methods for solving the problem, and 3) baseline method. As such it provides a unique opportunity for the principled investigation into various aspects of autonomous vehicle deployment.

DISK-CSV: Distilling Interpretable Semantic Knowledge with a Class Semantic Vector

Proceedings of the 16th Conference of the European Chapter of the Association for Computational Linguistics: Main Volume, 2021

Neural networks (NN) applied to natural language processing (NLP) are becoming deeper and more co... more Neural networks (NN) applied to natural language processing (NLP) are becoming deeper and more complex, making them increasingly difficult to understand and interpret. Even in applications of limited scope on fixed data, the creation of these complex "black-boxes" creates substantial challenges for debugging, understanding, and generalization. But rapid development in this field has now lead to building more straightforward and interpretable models. We propose a new technique (DISK-CSV) to distill knowledge concurrently from any neural network architecture for text classification, captured as a lightweight interpretable/explainable classifier. Across multiple datasets, our approach achieves better performance than the target black-box. In addition, our approach provides better explanations than existing techniques.

Depression symptoms modelling from social media text: an LLM driven semi-supervised learning approach

Language Resources and Evaluation

A fundamental component of user-level social media language based clinical depression modelling i... more A fundamental component of user-level social media language based clinical depression modelling is depression symptoms detection (DSD). Unfortunately, there does not exist any DSD dataset that reflects both the clinical insights and the distribution of depression symptoms from the samples of self-disclosed depressed population. In our work, we describe a semi-supervised learning (SSL) framework which uses an initial supervised learning model that leverages (1) a state-of-the-art large mental health forum text pre-trained language model further fine-tuned on a clinician annotated DSD dataset, (2) a Zero-Shot learning model for DSD, and couples them together to harvest depression symptoms related samples from our large self-curated depressive tweets repository (DTR). Our clinician annotated dataset is the largest of its kind. Furthermore, DTR is created from the samples of tweets in self-disclosed depressed users Twitter timeline from two datasets, including one of the largest benchma...

DISK-CSV: Distilling Interpretable Semantic Knowledge with a Class Semantic Vector

Neural networks (NN) applied to natural language processing (NLP) are becoming deeper and more co... more Neural networks (NN) applied to natural language processing (NLP) are becoming deeper and more complex, making them increasingly difficult to understand and interpret. Even in applications of limited scope on fixed data, the creation of these complex "black-boxes" creates substantial challenges for debugging, understanding, and generalization. But rapid development in this field has now lead to building more straightforward and interpretable models. We propose a new technique (DISK-CSV) to distill knowledge concurrently from any neural network architecture for text classification, captured as a lightweight interpretable/explainable classifier. Across multiple datasets, our approach achieves better performance than the target black-box. In addition, our approach provides better explanations than existing techniques.

Towards Integrative Machine Learning and Knowledge Extraction

Lecture Notes in Computer Science, 2017

The use of general descriptive names, registered names, trademarks, service marks, etc. in this p... more The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, express or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

A Multi-Task Learning Framework for Abstractive Text Summarization

Proceedings of the ... AAAI Conference on Artificial Intelligence, Jul 17, 2019

We propose a Multi-task learning approach for Abstractive Text Summarization (MATS), motivated by... more We propose a Multi-task learning approach for Abstractive Text Summarization (MATS), motivated by the fact that humans have no difficulty performing such task because they have the capabilities of multiple domains. Specifically, MATS consists of three components: (i) a text categorization model that learns rich category-specific text representations using a bi-LSTM encoder; (ii) a syntax labeling model that learns to improve the syntax-aware LSTM decoder; and (iii) an abstractive text summarization model that shares its encoder and decoder with the text categorization and the syntax labeling tasks, respectively. In particular, the abstractive text summarization model enjoys significant benefit from the additional text categorization and syntax knowledge. Our experimental results show that MATS outperforms the competitors. 1

Neural Networks with Feature Attribution and Contrastive Explanations

Springer eBooks, 2023

A negation detection assessment of GPTs: analysis with the xNot360 dataset

arXiv (Cornell University), Jun 28, 2023

Negation is a fundamental aspect of natural language, playing a critical role in communication an... more Negation is a fundamental aspect of natural language, playing a critical role in communication and comprehension. Our study assesses the negation detection performance of Generative Pretrained Transformer (GPT) models, specifically GPT-2, GPT-3, GPT-3.5, and GPT-4. We focus on the identification of negation in natural language using a zero-shot prediction approach applied to our custom xNot360 dataset. Our approach examines sentence pairs labeled to indicate whether the second sentence negates the first. Our findings expose a considerable performance disparity among the GPT models, with GPT-4 surpassing its counterparts and GPT-3.5 displaying a marked performance reduction. The overall proficiency of the GPT models in negation detection remains relatively modest, indicating that this task pushes the boundaries of their natural language understanding capabilities. We not only highlight the constraints of GPT models in handling negation but also emphasize the importance of logical reliability in high-stakes domains such as healthcare, science, and law.

Explaining Autonomous Driving Actions with Visual Question Answering

arXiv (Cornell University), Jul 19, 2023

The end-to-end learning ability of self-driving vehicles has achieved significant milestones over... more The end-to-end learning ability of self-driving vehicles has achieved significant milestones over the last decade owing to rapid advances in deep learning and computer vision algorithms. However, as autonomous driving technology is a safety-critical application of artificial intelligence (AI), road accidents and established regulatory principles necessitate the need for the explainability of intelligent action choices for self-driving vehicles. To facilitate interpretability of decisionmaking in autonomous driving, we present a Visual Question Answering (VQA) framework, which explains driving actions with question-answering-based causal reasoning. To do so, we first collect driving videos in a simulation environment using reinforcement learning (RL) and extract consecutive frames from this log data uniformly for five selected action categories. Further, we manually annotate the extracted frames using question-answer pairs as justifications for the actions chosen in each scenario. Finally, we evaluate the correctness of the VQA-predicted answers for actions on unseen driving scenes. The empirical results suggest that the VQA mechanism can provide support to interpret real-time decisions of autonomous vehicles and help enhance overall driving safety.

How well do SOTA legal reasoning models support abductive reasoning?

arXiv (Cornell University), Apr 13, 2023

We examine how well the state-of-the-art (SOTA) models used in legal reasoning support abductive ... more We examine how well the state-of-the-art (SOTA) models used in legal reasoning support abductive reasoning tasks. Abductive reasoning is a form of logical inference in which a hypothesis is formulated from a set of observations, and that hypothesis is used to explain the observations. The ability to formulate such hypotheses is important for lawyers and legal scholars as it helps them articulate logical arguments, interpret laws, and develop legal theories. Our motivation is to consider the belief that deep learning models, especially large language models (LLMs), will soon replace lawyers because they perform well on tasks related to legal text processing. But to do so, we believe, requires some form of abductive hypothesis formation. In other words, while LLMs become more popular and powerful, we want to investigate their capacity for abductive reasoning. To pursue this goal, we start by building a logic-augmented dataset for abductive reasoning with 498,697 samples and then use it to evaluate the performance of a SOTA model in the legal field. Our experimental results show that although these models can perform well on tasks related to some aspects of legal text processing, they still fall short in supporting abductive reasoning tasks.

Current Optical Sensing Applications in Seeds Vigor Determination

Agronomy, Apr 20, 2023

Advances in optical sensing technology have led to new approaches to monitoring and determining c... more Advances in optical sensing technology have led to new approaches to monitoring and determining crop seed vigor. In order to improve crop performance to secure reliable yield and food supply, calibrating seed vigor, purity, germination rate, and clarity is very critical to the future of the agriculture/horticulture industry. Traditional methods of seed vigor determination are lengthy in process, labor intensive, and sometimes inaccurate, which can lead to false yield prediction and faulty decision-making. Optical sensing technology offers rapid, accurate, and non-destructive calibration methods to help the industry develop accurate decisions for seed usage and agronomic evaluation. In this review, we hope to provide a summary of current research in the optical sensing technology used in seed vigor assessments.

Semantic-Based Classification of Relevant Case Law

New Frontiers in Artificial Intelligence

Basic and Depression Specific Emotions Identification in Tweets: Multi-label Classification Experiments

Lecture Notes in Computer Science, 2023

xxAI - Beyond Explainable Artificial Intelligence

xxAI - Beyond Explainable AI

The success of statistical machine learning from big data, especially of deep learning, has made ... more The success of statistical machine learning from big data, especially of deep learning, has made artificial intelligence (AI) very popular. Unfortunately, especially with the most successful methods, the results are very difficult to comprehend by human experts. The application of AI in areas that impact human life (e.g., agriculture, climate, forestry, health, etc.) has therefore led to an demand for trust, which can be fostered if the methods can be interpreted and thus explained to humans. The research field of explainable artificial intelligence (XAI) provides the necessary foundations and methods. Historically, XAI has focused on the development of methods to explain the decisions and internal mechanisms of complex AI systems, with much initial research concentrating on explaining how convolutional neural networks produce image classification predictions by producing visualizations which highlight what input patterns are most influential in activating hidden units, or are most ...

A Sketch of a Theory of Visualization

Proceedings of the 5th International Conference on Information Visualization Theory and Applications, 2014

A picture results from a possibly multi-layer transformation of data to a visual vocabulary in wh... more A picture results from a possibly multi-layer transformation of data to a visual vocabulary in which humans can draw inferences about the original data. The goal of this visualization process is to expose relationships amongst the data that are otherwise difficult to find, or only emerge by the process of the transformation. In case of the former kind of inference (confirming a relationship that did exist but was not obvious), visualization provides a kind of inferential amplifying effect. In the case of the latter (exposing new data relationships), visualization provides an inductive mechanism to create hypotheses not manifest in the original data. In this regard, the creation of pictures from data is about data compression, which is naturally a kind of machine learning. Just as statistical concepts like average and standard deviation provide a measure on properties of a set of numbers, so too does visualization provide a kind of "measure" on data compressed to a visual vocabulary presented as a picture. Our position is that visualization is about the (potentially multi-step, multilayered) transformation of data to pictures, and that ever such transformation must make choices about what kinds of relations to preserve, and what kinds of data artifacts to avoid in each such transformation. Like a chain of formal inference, conclusions following from the end result (the picture) are determined by what each transformation in the inference chain is intended to accomplish. We argue that the visualization of large data sets, too large to inspect directly, requires a rigorous theory of how to transform data to pictures, so that the scientists as observers can be assured that inferences drawn from the pictures are either confirmable in the detailed data, or at least plausible hypotheses which can be further pursued by seeking further data (evidence).

Preface of Special Issue on 8th Competition on Legal Information of Extraction and Entailment (COLIEE 2021)

The Review of Socionetwork Strategies, 2022

The Review of Socionetwork Strategies 1 3 develop and extend the case law data for COLIEE, and to... more The Review of Socionetwork Strategies 1 3 develop and extend the case law data for COLIEE, and to Young Yik Rhim of Intellicon in Seoul, who has been our advocate since the beginning of COLIEE. In addition, a number of Japanese colleagues (in addition to the organizing team of Ken Satoh, Yoshinobu Kano, and Masaharu Yoshioka) have contributed to the extension and curation of the statute law data for the COLIEE competition.

Explain Yourself – A Semantic Stack for Artificial Intelligence

The Application of Text Entailment Techniques in COLIEE 2020

New Frontiers in Artificial Intelligence, 2021

Facial expression recognition using SVM classification on mic-macro patterns

2017 IEEE International Conference on Image Processing (ICIP), 2017

The identification of facial expressions is a fundamental topic in the area of human computer int... more The identification of facial expressions is a fundamental topic in the area of human computer interaction and pattern recognition. The research has gained significant attention in recent years. However many challenges still exist. This is because an individual might display different expressions at different times for the same mood. Expressions can also be influenced by health. Our proposed framework aims to capture unique information related to expressions from salient patches. We extract representative feature patterns at both micro and macro levels within a pixel-patch, and use a support vector machine (SVM) classifier to label expressions. Our experimental results using the Japanese facial expression (JAFEE) and Cohn-Kanade (CK) datasets achieve high recognition rate and efficient computation time, outperforming existing work.

Overview of COLIEE 2017

We present the evaluation of the legal question answering Competition on Legal Information Extrac... more We present the evaluation of the legal question answering Competition on Legal Information Extraction/Entailment (COLIEE) 2017. The COLIEE 2017 Task consists of two sub-Tasks: legal information retrieval (Task 1), and recognizing entailment between articles and queries (Task 2). Participation was open to any group based on any approach, and the tasks attracted 10 teams. We received 9 submissions to Task 1 (for a total of 17 runs), and 8 submissions to Task 2 (for a total of 20 runs).