Rare Gems: Finding Lottery Tickets At Initialization,
2022
University of Wisconsin-Madison
Rare Gems: Finding Lottery Tickets At Initialization, Kartik Sreenivasan, Jy Yong Sohn, Liu Yang, Matthew Grinde, Alliot Nagle, Hongyi Wang, Eric Xing, Kangwook Lee, Dimitris Papailiopoulos
Machine Learning Faculty Publications
Large neural networks can be pruned to a small fraction of their original size, with little loss in accuracy, by following a time-consuming “train, prune, re-train” approach. Frankle & Carbin [9] conjecture that we can avoid this by training lottery tickets, i.e., special sparse subnetworks found at initialization, that can be trained to high accuracy. However, a subsequent line of work [11, 41] presents concrete evidence that current algorithms for finding trainable networks at initialization, fail simple baseline comparisons, e.g., against training random sparse subnetworks. Finding lottery tickets that train to better accuracy compared to simple baselines remains an open …
Unpaired Image-To-Image Translation With Density Changing Regularization,
2022
Carnegie Mellon University
Unpaired Image-To-Image Translation With Density Changing Regularization, Shaoan Xie, Qirong Ho, Kun Zhang
Machine Learning Faculty Publications
Unpaired image-to-image translation aims to translate an input image to another domain such that the output image looks like an image from another domain while important semantic information are preserved. Inferring the optimal mapping with unpaired data is impossible without making any assumptions. In this paper, we make a density changing assumption where image patches of high probability density should be mapped to patches of high probability density in another domain. Then we propose an efficient way to enforce this assumption: we train the flows as density estimators and penalize the variance of density changes. Despite its simplicity, our method …
The Role Of Generative Adversarial Networks In Bioimage Analysis And Computational Diagnostics.,
2022
University of Louisville
The Role Of Generative Adversarial Networks In Bioimage Analysis And Computational Diagnostics., Ahmed Naglah
Electronic Theses and Dissertations
Computational technologies can contribute to the modeling and simulation of the biological environments and activities towards achieving better interpretations, analysis, and understanding. With the emergence of digital pathology, we can observe an increasing demand for more innovative, effective, and efficient computational models. Under the umbrella of artificial intelligence, deep learning mimics the brain’s way in learn complex relationships through data and experiences. In the field of bioimage analysis, models usually comprise discriminative approaches such as classification and segmentation tasks. In this thesis, we study how we can use generative AI models to improve bioimage analysis tasks using Generative Adversarial Networks …
Identity Term Sampling For Measuring Gender Bias In Training Data,
2022
Technological University Dublin
Identity Term Sampling For Measuring Gender Bias In Training Data, Nasim Sobhani, Sarah Jane Delany
Conference Papers
Predictions from machine learning models can reflect biases in the data on which they are trained. Gender bias has been identified in natural language processing systems such as those used for recruitment. The development of approaches to mitigate gender bias in training data typically need to be able to isolate the effect of gender on the output to see the impact of gender. While it is possible to isolate and identify gender for some types of training data, e.g. CVs in recruitment, for most textual corpora there is no obvious gender label. This paper proposes a general approach to measure …
A Logistic Regression And Linear Programming Approach For Multi-Skill Staffing Optimization In Call Centers,
2022
Singapore Management University
A Logistic Regression And Linear Programming Approach For Multi-Skill Staffing Optimization In Call Centers, Thuy Anh Ta, Tien Mai, Fabian Bastin, Pierre L'Ecuyer
Research Collection School Of Computing and Information Systems
We study a staffing optimization problem in multi-skill call centers. The objective is to minimize the total cost of agents under some quality of service (QoS) constraints. The key challenge lies in the fact that the QoS functions have no closed-form and need to be approximated by simulation. In this paper we propose a new way to approximate the QoS functions by logistic functions and design a new algorithm that combines logistic regression, cut generations and logistic-based local search to efficiently find good staffing solutions. We report computational results using examples up to 65 call types and 89 agent groups …
Divide-And-Conquer Distributed Learning: Privacy-Preserving Offloading Of Neural Network Computations,
2022
University of Arkansas, Fayetteville
Divide-And-Conquer Distributed Learning: Privacy-Preserving Offloading Of Neural Network Computations, Lewis C.L. Brown
Graduate Theses and Dissertations
Machine learning has become a highly utilized technology to perform decision making on high dimensional data. As dataset sizes have become increasingly large so too have the neural networks to learn the complex patterns hidden within. This expansion has continued to the degree that it may be infeasible to train a model from a singular device due to computational or memory limitations of underlying hardware. Purpose built computing clusters for training large models are commonplace while access to networks of heterogeneous devices is still typically more accessible. In addition, with the rise of 5G networks, computation at the edge becoming …
Performance Enhancement Of Hyperspectral Semantic Segmentation Leveraging Ensemble Networks,
2022
University of Maine
Performance Enhancement Of Hyperspectral Semantic Segmentation Leveraging Ensemble Networks, Nicholas Soucy
Electronic Theses and Dissertations
Hyperspectral image (HSI) semantic segmentation is a growing field within computer vision, machine learning, and forestry. Due to the separate nature of these communities, research applying deep learning techniques to ground-type semantic segmentation needs improvement, along with working to bring the research and expectations of these three communities together. Semantic segmentation consists of classifying individual pixels within the image based on the features present. Many issues need to be resolved in HSI semantic segmentation including data preprocessing, feature reduction, semantic segmentation techniques, and adversarial training. In this thesis, we tackle these challenges by employing ensemble methods for HSI semantic segmentation. …
A Protocol To Build Trust With Black Box Models,
2022
University of Wisconsin-Milwaukee
A Protocol To Build Trust With Black Box Models, Timothy K. Thielke
Theses and Dissertations
Data scientists are more widely using artificial intelligence and machine learning (ML) algorithms today despite the general mistrust associated with them due to the lack of contextual understanding of the domain occurring within the algorithm. Of the many types of ML algorithms, those that use non-linear activation functions are especially regarded with suspicion because of the lack of transparency and intuitive understanding of what is occurring within the black box of the algorithm. In this thesis, we set out to create a protocol to delve into the black box of an ML algorithm set to predict synoptic severe weather patterns …
End-To-End Hierarchical Reinforcement Learning With Integrated Subgoal Discovery,
2022
Singapore Management University
End-To-End Hierarchical Reinforcement Learning With Integrated Subgoal Discovery, Shubham Pateria, Budhitama Subagdja, Ah-Hwee Tan, Chai Quek
Research Collection School Of Computing and Information Systems
Hierarchical reinforcement learning (HRL) is a promising approach to perform long-horizon goal-reaching tasks by decomposing the goals into subgoals. In a holistic HRL paradigm, an agent must autonomously discover such subgoals and also learn a hierarchy of policies that uses them to reach the goals. Recently introduced end-to-end HRL methods accomplish this by using the higher-level policy in the hierarchy to directly search the useful subgoals in a continuous subgoal space. However, learning such a policy may be challenging when the subgoal space is large. We propose integrated discovery of salient subgoals (LIDOSS), an end-to-end HRL method with an integrated …
Data-Driven Deep Learning-Based Analysis On Thz Imaging,
2022
University of Arkansas, Fayetteville
Data-Driven Deep Learning-Based Analysis On Thz Imaging, Haoyan Liu
Graduate Theses and Dissertations
Breast cancer affects about 12.5% of women population in the United States. Surgical operations are often needed post diagnosis. Breast conserving surgery can help remove malignant tumors while maximizing the remaining healthy tissues. Due to lacking effective real-time tumor analysis tools and a unified operation standard, re-excision rate could be higher than 30% among breast conserving surgery patients. This results in significant physical, physiological, and financial burdens to those patients. This work designs deep learning-based segmentation algorithms that detect tissue type in excised tissues using pulsed THz technology. This work evaluates the algorithms for tissue type classification task among freshly …
Wildfire Spread Prediction Using Attention Mechanisms In U-Net,
2022
Calpoly
Wildfire Spread Prediction Using Attention Mechanisms In U-Net, Kamen Haresh Shah, Kamen Haresh Shah
Master's Theses
An investigation into using attention mechanisms for better feature extraction in wildfire spread prediction models. This research examines the U-net architecture to achieve image segmentation, a process that partitions images by classifying pixels into one of two classes. The deep learning models explored in this research integrate modern deep learning architectures, and techniques used to optimize them. The models are trained on 12 distinct observational variables derived from the Google Earth Engine catalog. Evaluation is conducted with accuracy, Dice coefficient score, ROC-AUC, and F1-score. This research concludes that when augmenting U-net with attention mechanisms, the attention component improves feature suppression …
Greener: Graph Neural Networks For News Media Profiling,
2022
Sofia University St. Kliment Ohridski
Greener: Graph Neural Networks For News Media Profiling, Panayot Panayotov, Utsav Shukla, Husrev T. Sencar, Mohamed Nabeel, Preslav Nakov
Natural Language Processing Faculty Publications
We study the problem of profiling news media on the Web with respect to their factuality of reporting and bias. This is an important but under-studied problem related to disinformation and “fake news” detection, but it addresses the issue at a coarser granularity compared to looking at an individual article or an individual claim. This is useful as it allows to profile entire media outlets in advance. Unlike previous work, which has focused primarily on text (e.g., on the articles published by the target website, or on the textual description in their social media profiles or in Wikipedia), here we …
Motion Planning Under Uncertainties,
2022
University at Albany, State University of New York
Motion Planning Under Uncertainties, Sourav Dutta
Legacy Theses & Dissertations (2009 - 2024)
A robot is an agent that can bring some changes to the environment around it. Motion planning is the problem of carrying out specialized tasks by a robot by either moving itself or some other object (usually called \textit{payload}) from one place to another. In a real-world scenario, a robot is faced with constraints such as momentum, friction, sensor inaccuracies, etc., that can affect its decision-making while performing specialized tasks. These constraints are identified as uncertainties, and successful planning involves making provisions for such uncertainties. In this work, we present methods like stochastic processes, sequential inference, and pattern recognition to …
Probabilistic Forecasting Of Winter Mixed Precipitation Types In New York State Utilizing A Random Forest,
2022
University at Albany, State University of New York
Probabilistic Forecasting Of Winter Mixed Precipitation Types In New York State Utilizing A Random Forest, Brian Chandler Filipiak
Legacy Theses & Dissertations (2009 - 2024)
Operational forecasters face a plethora of challenges when making a forecast; they must consider multiple data sources ranging from radar and satellites to surface and upper air observations, to numerical weather prediction output. Forecasts must be done in a limited window of time, which adds an additional layer of difficulty to the task. These challenges are exacerbated by winter mixed precipitation events where slight differences in thermodynamic profiles or changes in terrain create different precipitation types across small areas. In addition to being difficult to forecast, mixed precipitation events can have large-scale impacts on our society.
Development Of Nucleic Acid Diagnostics For Targeted And Non-Targeted Biosensing,
2022
University at Albany, State University of New York
Development Of Nucleic Acid Diagnostics For Targeted And Non-Targeted Biosensing, Christopher William Smith
Legacy Theses & Dissertations (2009 - 2024)
The field of nucleic acid technology is rapidly expanding with new impactful discoveriesbeing made each year. Starting from the discovery of the double-helix structure, cloning, gene editing, polymerase chain reaction (PCR), CRISPR technology, and even the late mRNA vaccines; nucleic acid technology is at the forefront of improving medicine. Nucleic acid technology is extremely versatile due to its easy programmability, automated cheap synthesis, and even its catalog for numerous chemical modifications that can be used to alter structure stability. For example, the number of permutations that can be made with DNA just by altering the code for adenine (A), cytosine …
A Unified Dialogue User Simulator For Few-Shot Data Augmentation,
2022
Tsinghua University
A Unified Dialogue User Simulator For Few-Shot Data Augmentation, Dazhen Wan, Zheng Zhang, Qi Zhu, Lizi Liao, Minlie Huang
Research Collection School Of Computing and Information Systems
Pre-trained language models have shown superior performance in task-oriented dialogues. However, existing datasets are on limited scales, which cannot support large-scale pre-training. Fortunately, various data augmentation methods have been developed to augment largescale task-oriented dialogue corpora. However, they heavily rely on annotated data in the target domain, which require a tremendous amount of data collection and human labeling work. In this paper, we build a unified dialogue user simulation model by pre-training on several publicly available datasets. The model can then be tuned on a target domain with fewshot data. The experiments on a target dataset across multiple domains show …
Aspect Sentiment Triplet Extraction Incorporating Syntactic Constituency Parsing Tree And Commonsense Knowledge Graph,
2022
Singapore Management University
Aspect Sentiment Triplet Extraction Incorporating Syntactic Constituency Parsing Tree And Commonsense Knowledge Graph, Zhenda Hu, Zhaoxia Wang, Yinglin Wang, Ah-Hwee Tan
Research Collection School Of Computing and Information Systems
The aspect sentiment triplet extraction (ASTE) task aims to extract the target term and the opinion term, and simultaneously identify the sentiment polarity of target-opinion pairs from the given sentences. While syntactic constituency information and commonsense knowledge are both important and valuable for the ASTE task, only a few studies have explored how to integrate them via flexible graph convolutional networks (GCNs) for this task. To address this gap, this paper proposes a novel end-to-end model, namely GCN-EGTS, which is an enhanced Grid Tagging Scheme (GTS) for ASTE leveraging syntactic constituency parsing tree and a commonsense knowledge graph based on …
Prompting For Multimodal Hateful Meme Classification,
2022
Singapore Management University
Prompting For Multimodal Hateful Meme Classification, Rui Cao, Roy Ka-Wei Lee, Wen-Haw Chong, Jing Jiang
Research Collection School Of Computing and Information Systems
Hateful meme classification is a challenging multimodal task that requires complex reasoning and contextual background knowledge. Ideally, we could leverage an explicit external knowledge base to supplement contextual and cultural information in hateful memes. However, there is no known explicit external knowledge base that could provide such hate speech contextual information. To address this gap, we propose PromptHate, a simple yet effective prompt-based model that prompts pre-trained language models (PLMs) for hateful meme classification. Specifically, we construct simple prompts and provide a few in-context examples to exploit the implicit knowledge in the pretrained RoBERTa language model for hateful meme classification. …
Expressiveness Of Real-Time Motion Captured Avatars Influences Perceived Animation Realism And Perceived Quality Of Social Interaction In Virtual Reality,
2022
Edith Cowan University
Expressiveness Of Real-Time Motion Captured Avatars Influences Perceived Animation Realism And Perceived Quality Of Social Interaction In Virtual Reality, Alan D. Fraser, Isabella Branson, Ross C. Hollett, Craig P. Speelman, Shane L. Rogers
Research outputs 2022 to 2026
Using motion capture to enhance the realism of social interaction in virtual reality (VR) is growing in popularity. However, the impact of different levels of avatar expressiveness on the user experience is not well understood. In the present study we manipulated levels of face and body expressiveness of avatars while investigating participant perceptions of animation realism and interaction quality when disclosing positive and negative experiences in VR. Moderate positive associations were observed between perceptions of animation realism and interaction quality. Post-experiment questions revealed that many of our participants (approximately 40 %) indicated the avatar with the highest face and body …
Conversation Disentanglement With Bi-Level Contrastive Learning,
2022
National University of Singapore
Conversation Disentanglement With Bi-Level Contrastive Learning, Chengyu Huang, Zheng Zhang, Hao Fei, Lizi Liao
Research Collection School Of Computing and Information Systems
Conversation disentanglement aims to group utterances into detached sessions, which is a fundamental task in processing multi-party conversations. Existing methods have two main drawbacks. First, they overemphasize pairwise utterance relations but pay inadequate attention to the utterance-to-context relation modeling. Second, a huge amount of human annotated data is required for training, which is expensive to obtain in practice. To address these issues, we propose a general disentangle model based on bi-level contrastive learning. It brings closer utterances in the same session while encourages each utterance to be near its clustered session prototypes in the representation space. Unlike existing approaches, our …
