YorkSpace
YorkSpace is York University's Institutional Repository. It supports York University's Senate Policy on Open Access by providing York community members with a place to preserve their research online in an institutional context.
Communities in YorkSpace
Select a community to browse its collections.
- Previously Faculty of Environmental Studies (FES)
- The Global Labour Research Centre (GLRC) engages in the study of work, employment and labour in the context of a constantly changing global economy.
- Lives Outside the Lines: a Symposium in Honour of Marlene Kadar
- Used only for SWORD Deposit by Adminstrator
- Welcome to WILAA, a gathering place for materials related to research projects that explore work-integrated learning and disability-related accessibility and accommodations.
Recent Submissions
Exploiting Reward Machines with Deep Reinforcement Learning in Continuous Action Domains
(Springer Cham, 2023-09-07) Haolin Sun; Lesperance, Yves
In this paper, we address the challenges of non-Markovian rewards and learning efficiency in deep reinforcement learning (DRL) in continuous action domains by exploiting reward machines (RMs) and counterfactual experiences for reward machines (CRM). RM and CRM were proposed by Toro Icarte et al. A reward machine can decompose a task, convey its high-level structure to an agent, and support certain non-Markovian task specifications. In this paper, we integrate state-of-the-art DRL algorithms with RMs to enhance learning efficiency. Our experimental results demonstrate that Soft Actor-Critic with counterfactual experiences for RMs (SAC-CRM) facilitates faster learning of better policies, while Deep Deterministic Policy Gradient with counterfactual experiences for RMs (DDPG-CRM) is slower, achieves lower rewards, but is more stable. Option-based Hierarchical Reinforcement Learning for reward machines (HRM) and Twin Delayed Deep Deterministic (TD3) with CRM generally underperform compared to SAC-CRM and DDPG-CRM. This work contributes to the ongoing development of more efficient and robust DRL approaches by leveraging the potential of RMs in practical problem-solving scenarios.
Characterization Of A Cysteine Protease From Phytolacca Americana And Its Association With Pokeweed Antiviral Protein
(2024-10-28) Audet, Annabelle; Hudak, Kathi
The plant apoplast, an essential extracellular space, contains diverse proteins crucial for plant defence. This study focuses on Phytolacca americana (American pokeweed), examining the interaction between the ribosome inactivating protein Pokeweed Antiviral Protein (PAP) and a putative papain-like cysteine protease, Phytolacca americana cysteine protease 1 (PaCP1). Bioinformatic predictions identified PaCP1 as a papain-like cysteine protease with conserved structural features, while enzymatic assays confirmed its functionality. Yeast-two hybrid assays validated the interaction between PAP and PaCP1, and localization studies indicated their extracellular co-localization in the apoplast. Enzymatic assays further demonstrated PaCP1's ability to cleave PAP, with mass spectrometry identifying the resulting degradation products. Differential expression studies under salicylic acid and Flg22 treatment revealed different patterns of expression for both PAP and PaCP1. This research enhances our understanding of PAP and its protein interactions as well as its potential role in the pokeweed apoplast.
Effects Of Cold-Water Immersion On Post-Exercise Skeletal Muscle Recovery Following Sprint-Interval Exercise
(2024-10-28) Richards, Andrew Jason; Cheng, Arthur
Cold-water immersion has emerged as a popular post-exercise recovery intervention for athletes and avid exercisers. However, little evidence exists to support its widespread use, especially following high-intensity interval exercise (HIIE). Therefore, to investigate the use of CWI following HIIE, 12 participants participated in a randomized cross-over study involving repeated all-out contractions of the ankle dorsiflexor muscles followed by CWI or room temperature rest (RT). During a 24-h recovery period, neuromuscular function, intramuscular temperature, and next-day HIIE were performed. The results of the study showed that CWI impaired maximal tetanic force for up to 3-h, whereas immediate recovery occurred following RT. In addition, there was no difference in next-day HIIE performance between the two recovery interventions. Thus, CWI offers no substantial benefit as an effective post-exercise recovery intervention following HIIE.
Probing Human Visual Strategies Using Interpretability Methods for Artificial Neural Networks
(2024-10-28) Kashef Alghetaa, Yousif Khalid Faeq; Kar, Kohitij
Unraveling human visual strategies during object recognition remains a challenge in vision science. Existing psychophysical methods used to investigate these strategies are limited in accurately interpreting human decisions. Recently, artificial neural network (ANN) models, which show remarkable similarities to human vision, provide a window into human visual strategies. However, inconsistencies among different techniques hinder the use of explainable AI (XAI) methods to interpret ANN decision-making. Here, we first develop and validate a novel surrogate method, in silico, using behavioral probes in ANNs with explanation-masked images to address these challenges. Finally, by identifying the XAI method and ANN with the highest human alignment, we provide a working hypothesis and an effective approach to explain human visual strategies during object recognition -- a framework relevant to many other behaviors.
Progressive Hierarchical Classification For Multi-Category Image Classification
(2024-10-28) Kuo, Te-Chuan; Chen, Stephen
This thesis evaluates a hierarchical classification model applied to the CIFAR-10 dataset, focusing on addressing the limitations of existing methods, which often struggle with (i) overlapping features and (ii) poor interpretability of classification decisions. The hierarchical model was implemented to mitigate these issues by refining classification through a multi-stage process that narrows the focus progressively.
Our hierarchical approach has demonstrated its ability to focus on distinguishing features critical to specific classes and groups/pairs. Furthermore, the hierarchical models provide enhanced transparency over the baseline model by allowing a granular examination of classification performance across multiple stages.