Physical Sciences and Mathematics Commons^™

Ur2m: Uncertainty And Resource-Aware Event Detection On Microcontrollers, Hong Jia, Young D. Kwon, Dong Ma, Nhat Pham, Lorena Qendro, Tam Vu, Cecilia Mascolo

Research Collection School Of Computing and Information Systems

Traditional machine learning techniques are prone to generating inaccurate predictions when confronted with shifts in the distribution of data between the training and testing phases. This vulnerability can lead to severe consequences, especially in applications such as mobile healthcare. Uncertainty estimation has the potential to mitigate this issue by assessing the reliability of a model's output. However, existing uncertainty estimation techniques often require substantial computational resources and memory, making them impractical for implementation on microcontrollers (MCUs). This limitation hinders the feasibility of many important on-device wearable event detection (WED) applications, such as heart attack detection. In this paper, we present …

Affinity Uncertainty-Based Hard Negative Mining In Graph Contrastive Learning, Chaoxi Niu, Guansong Pang, Ling Chen

Research Collection School Of Computing and Information Systems

Hard negative mining has shown effective in enhancing self-supervised contrastive learning (CL) on diverse data types, including graph CL (GCL). The existing hardness-aware CL methods typically treat negative instances that are most similar to the anchor instance as hard negatives, which helps improve the CL performance, especially on image data. However, this approach often fails to identify the hard negatives but leads to many false negatives on graph data. This is mainly due to that the learned graph representations are not sufficiently discriminative due to oversmooth representations and/or non-independent and identically distributed (non-i.i.d.) issues in graph data. To tackle this …

Active Discovering New Slots For Task-Oriented Conversation, Yuxia Wu, Tianhao Dai, Zhedong Zheng, Lizi Liao

Research Collection School Of Computing and Information Systems

Existing task-oriented conversational systems heavily rely on domain ontologies with pre-defined slots and candidate values. In practical settings, these prerequisites are hard to meet, due to the emerging new user requirements and ever-changing scenarios. To mitigate these issues for better interaction performance, there are efforts working towards detecting out-of-vocabulary values or discovering new slots under unsupervised or semi-supervised learning paradigms. However, overemphasizing on the conversation data patterns alone induces these methods to yield noisy and arbitrary slot results. To facilitate the pragmatic utility, real-world systems tend to provide a stringent amount of human labeling quota, which offers an authoritative way …

Designing Large-Scale Intelligent Collaborative Platform For Freight Forwarders, Pang Jin Tan, Shih-Fen Cheng, Richard Chen

Research Collection School Of Computing and Information Systems

In this paper, we propose to design a large-scale intelligent collaborative platform for freight forwarders. This platform is based on a mathematical programming formulation and an efficient solution approach. Forwarders are middlemen who procure container capacities from carriers and sell them to shippers to serve their transport requests. However, due to demand uncertainty, they often either over-procure or under-procure capacities. We address this with our proposed platform where forwarders can collaborate and share capacities, allowing one's transport requests to be potentially shipped on another forwarder's container. The result is lower total costs for all participating forwarders. The collaboration can be …

Clustering-Nn Based Cfo Estimation Using Random Access Preambles For 5g Non-Terrestrial Networks, Li Zhen, Luyao Cheng, Zheng Chu, Keping Yu, Pei Xiao, Mohsen Guizani

Machine Learning Faculty Publications

Non-terrestrial networks (NTNs) are expected to play a pivotal role in the future wireless ecosystem. Due to its high-dynamic characteristics, the accurate estimation and compensation of carrier frequency offset (CFO) are crucial for supporting 5G new radio (NR) enabled satellite direct access. With emphasis on ensuring reliable uplink synchronization, we propose a clustering-neural network based CFO estimation scheme by virtue of NR random access preambles. By leveraging the sparsity and regularity of input samples, the proposed scheme can achieve fast and precise prediction of CFOs, while establishing robustness against time uncertainty and channel variation within a satellite beam. Simulation results …

Optimizing Uncertainty Quantification Of Vision Transformers In Deep Learning On Novel Ai Architectures, Erik Pautsch, John Li, Silvio Rizzi, George K. Thiruvathukal, Maria Pantoja

Computer Science: Faculty Publications and Other Works

Deep Learning (DL) methods have shown substantial efficacy in computer vision (CV) and natural language processing (NLP). Despite their proficiency, the inconsistency in input data distributions can compromise prediction reliability. This study mitigates this issue by introducing uncertainty evaluations in DL models, thereby enhancing dependability through a distribution of predictions. Our focus lies on the Vision Transformer (ViT), a DL model that harmonizes both local and global behavior. We conduct extensive experiments on the ImageNet-1K dataset, a vast resource with over a million images across 1,000 categories. ViTs, while competitive, are vulnerable to adversarial attacks, making uncertainty estimation crucial for …

Optimized Uncertainty Estimation For Vision Transformers: Enhancing Adversarial Robustness And Performance Using Selective Classification, Erik Pautsch, John Li, Silvio Rizzi, George K. Thiruvathukal, Maria Pantoja

Computer Science: Faculty Publications and Other Works

Deep Learning models often exhibit undue confidence when encountering out-of-distribution (OOD) inputs, misclassifying with high confidence. The ideal outcome, in these cases, would be an "I do not know" verdict. We enhance the trustworthiness of our models through selective classification, allowing the model to abstain from making predictions when facing uncertainty. Rather than a singular prediction, the model offers a prediction distribution, enabling users to gauge the model’s trustworthiness and determine the need for human intervention. We assess uncertainty in two baseline models: a Convolutional Neural Network (CNN) and a Vision Transformer (ViT). By leveraging these uncertainty values, we minimize …

Uncertainty-Adjusted Inductive Matrix Completion With Graph Neural Networks, Petr Kasalicky, Antoine Ledent, Rodrigo Alves

Research Collection School Of Computing and Information Systems

We propose a robust recommender systems model which performs matrix completion and a ratings-wise uncertainty estimation jointly. Whilst the prediction module is purely based on an implicit low-rank assumption imposed via nuclear norm regularization, our loss function is augmented by an uncertainty estimation module which learns an anomaly score for each individual rating via a Graph Neural Network: data points deemed more anomalous by the GNN are downregulated in the loss function used to train the low-rank module. The whole model is trained in an end-to-end fashion, allowing the anomaly detection module to tap on the supervised information available in …

Multiclass Confidence And Localization Calibration For Object Detection, Bimsara Pathiraja, Malitha Gunawardhana, Muhammad Haris Khan

Computer Vision Faculty Publications

Albeit achieving high predictive accuracy across many challenging computer vision problems, recent studies suggest that deep neural networks (DNNs) tend to make over-confident predictions, rendering them poorly calibrated. Most of the existing attempts for improving DNN calibration are limited to classification tasks and restricted to calibrating in-domain predictions. Surprisingly, very little to no attempts have been made in studying the calibration of object detection methods, which occupy a pivotal space in vision-based security-sensitive, and safety-critical applications. In this paper, we propose a new train-time technique for calibrating modern object detection methods. It is capable of jointly calibrating multiclass confidence and …

Avoiding Starvation Of Arms In Restless Multi-Armed Bandit, Dexun Li, Pradeep Varakantham

Research Collection School Of Computing and Information Systems

Restless multi-armed bandits (RMAB) is a popular framework for optimizing performance with limited resources under uncertainty. It is an extremely useful model for monitoring beneficiaries (arms) and executing timely interventions using health workers (limited resources) to ensure optimal benefit in public health settings. For instance, RMAB has been used to track patients' health and monitor their adherence in tuberculosis settings, ensure pregnant mothers listen to automated calls about good pregnancy practices, etc. Due to the limited resources, typically certain individuals, communities, or regions are starved of interventions, which can potentially have a significant negative impact on the individual/community in the …

Performance Analysis Of Empirical Open-Circuit Voltage Modeling In Lithium-Ion Batteries, Part-3: Experimental Results, Prarthana Pillai, James Nguyen, Balakumar Balasingam

Computer Science Publications

This paper is the third part of a series of papers about empirical approaches to open circuit voltage (OCV) modeling of lithium-ion batteries. The first part of the series proposed models to quantify various sources of uncertainties in the OCV models; the second part of the series presented systematic data collection approaches to compute the uncertainties in the OCV to state of charge (SOC) models. This paper uses data collected from 28 OCV characterization experiments, performed according to the data collection plan presented in the second part, to compute and analyze three OCV uncertainty metrics: cell-to-cell variations, C-Rate error, and …

A Comparative Assessment Of Human Factors In Cybersecurity: Implications For Cyber Governance, Muhammad Umair Shah, Farkhund Iqbal, Umair Rehman, Patrick C.K. Hung

All Works

This paper provides an extensive overview of cybersecurity awareness in the young, educated, and technology-savvy population of the United Arab Emirates (UAE), compared to the United States of America (USA) for advancing the scholarship and practice of global cyber governance. We conducted comparative empirical studies to identify differences in specific human factors that affect cybersecurity behaviour in the UAE and the USA. In addition, we employed several control variables to observe reliable results. We used Hofstede’s theoretical framework on culture to advance our investigation. The results show that the targeted population in the UAE exhibits contrasting interpretations of cybersecurity awareness …

Learning Scenario Representation For Solving Two-Stage Stochastic Integer Programs, Yaoxin Wu, Wen Song, Zhiguang Cao, Jie Zhang

Research Collection School Of Computing and Information Systems

Many practical combinatorial optimization problems under uncertainty can be modeled as stochastic integer programs (SIPs), which are extremely challenging to solve due to the high complexity. To solve two-stage SIPs efficiently, we propose a conditional variational autoencoder (CVAE) based method to learn scenario representation for a class of SIP instances. Specifically, we design a graph convolutional network based encoder to embed each scenario with the deterministic part of its instance (i.e. context) into a low-dimensional latent space, from which a decoder reconstructs the scenario from its latent representation conditioned on the context. Such a design effectively captures the dependencies of …

Knowledge Graph Embedding By Normalizing Flows, Changyi Xiao, Xiangnan He, Yixin Cao

Research Collection School Of Computing and Information Systems

A key to knowledge graph embedding (KGE) is to choose a proper representation space, e.g., point-wise Euclidean space and complex vector space. In this paper, we propose a unified perspective of embedding and introduce uncertainty into KGE from the view of group theory. Our model can incorporate existing models (i.e., generality), ensure the computation is tractable (i.e., efficiency) and enjoy the expressive power of complex random variables (i.e., expressiveness). The core idea is that we embed entities/relations as elements of a symmetric group, i.e., permutations of a set. Permutations of different sets can reflect different properties of embedding. And the …

Jointly-Learnt Networks For Future Action Anticipation Via Self-Knowledge Distillation And Cycle Consistency, Md Moniruzzaman, Zhaozheng Yin, Zhihai He, Ming-Chuan Leu, Ruwen Qin

Mechanical and Aerospace Engineering Faculty Research & Creative Works

Future action anticipation aims to infer future actions from the observation of a small set of past video frames. In this paper, we propose a novel Jointly learnt Action Anticipation Network (J-AAN) via Self-Knowledge Distillation (Self-KD) and cycle consistency for future action anticipation. In contrast to the current state-of-the-art methods which anticipate the future actions either directly or recursively, our proposed J-AAN anticipates the future actions jointly in both direct and recursive ways. However, when dealing with future action anticipation, one important challenge to address is the future's uncertainty since multiple action sequences may come from or be followed by …

Communicating Uncertain Information From Deep Learning Models In Human Machine Teams, Harishankar V. Subramanian, Casey I. Canfield, Daniel Burton Shank, Luke Andrews, Cihan H. Dagli

Engineering Management and Systems Engineering Faculty Research & Creative Works

The role of human-machine teams in society is increasing, as big data and computing power explode. One popular approach to AI is deep learning, which is useful for classification, feature identification, and predictive modeling. However, deep learning models often suffer from inadequate transparency and poor explainability. One aspect of human systems integration is the design of interfaces that support human decision-making. AI models have multiple types of uncertainty embedded, which may be difficult for users to understand. Humans that use these tools need to understand how much they should trust the AI. This study evaluates one simple approach for communicating …

Co2vec: Embeddings Of Co-Ordered Networks Based On Mutual Reinforcement, Meng-Fen Chiang, Ee-Peng Lim, Wang-Chien Lee, Philips Kokoh Prasetyo

Research Collection School Of Computing and Information Systems

We study the problem of representation learning for multiple types of entities in a co-ordered network where order relations exist among entities of the same type, and association relations exist across entities of different types. The key challenge in learning co-ordered network embedding is to preserve order relations among entities of the same type while leveraging on the general consistency in order relations between different entity types. In this paper, we propose an embedding model, CO2Vec, that addresses this challenge using mutually reinforced order dependencies. Specifically, CO2Vec explores in-direct order dependencies as supplementary evidence to enhance order representation learning across …

Towards Characterizing Adversarial Defects Of Deep Learning Software From The Lens Of Uncertainty, Xiyue Zhang, Xiaofei Xie, Lei Ma, Xiaoning Du, Qiang Hu, Yang Liu, Jianjun Zhao, Meng Sun

Research Collection School Of Computing and Information Systems

Over the past decade, deep learning (DL) has been successfully applied to many industrial domain-specific tasks. However, the current state-of-the-art DL software still suffers from quality issues, which raises great concern especially in the context of safety- and security-critical scenarios. Adversarial examples (AEs) represent a typical and important type of defects needed to be urgently addressed, on which a DL software makes incorrect decisions. Such defects occur through either intentional attack or physical-world noise perceived by input sensors, potentially hindering further industry deployment. The intrinsic uncertainty nature of deep learning decisions can be a fundamental reason for its incorrect behavior. …

Adaptive Heuristics That (Could) Fit: Information Search And Communication Patterns In An Online Forum Of Investors Under Market Uncertainty, Niccolo Casnici, Marco Castellani, Flaminio Squazzoni, Manuela Testa, Pierpaolo Dondio

Articles

This article examines information-search heuristics and communication patterns in an online forum of investors during a period of market uncertainty. Global connections, real-time communication, and technological sophistication have created an unpredictable market environment. As such, investors try to deal with semantic, strategic, and operational uncertainty by following heuristics that reduce information redundancy. In this study, we have tried to find traces of cognitive communication heuristics in a large-scale data set including 8 years of online posts (2004–2012) for a forum of Italian investors. We identified various market volatility conditions on a daily basis to understand the influence of market uncertainty …

How To Deal With Uncertainties In Computing: From Probabilistic And Interval Uncertainty To Combination Of Different Approaches, With Applications To Engineering And Bioinformatics, Vladik Kreinovich

Departmental Technical Reports (CS)

Most data processing techniques traditionally used in scientific and engineering practice are statistical. These techniques are based on the assumption that we know the probability distributions of measurement errors etc.

In practice, often, we do not know the distributions, we only know the bound D on the measurement accuracy -- hence, after the get the measurement result X, the only information that we have about the actual (unknown) value x of the measured quantity is that $x$ belongs to the interval [X − D, X + D]. Techniques for data processing under such interval uncertainty are called interval computations; these …

Partitioning Uncertain Workloads, Freddy Chua, Bernardo A. Huberman

Research Collection School Of Computing and Information Systems

We present a method for determining the ratio of the tasks when breaking any complex workload in such a way that once the outputs from all tasks are joined, their full completion takes less time and exhibit smaller variance than when running on the undivided workload. To do that, we have to infer the capabilities of the processing unit executing the divided workloads or tasks. We propose a Bayesian Inference algorithm to infer the amount of time each task takes in a way that does not require prior knowledge on the processing unit capability. We demonstrate the effectiveness of this …

Experience Me! The Impact Of Content Sampling Strategies On The Marketing Of Digital Entertainment Goods, Ai Phuong Hoang, Robert J. Kauffman

Research Collection School Of Computing and Information Systems

Product sampling allows consumers to try out a small portion of a product for free. Uncertainty associated with consumption of information goods makes sampling useful for digital entertainment providers. Firms offer some programming for free to attract consumers to purchase a series of programs. We explore the effectiveness of content sampling for information goods using a dataset containing more than 17 million free previews and purchase observations on households from a digital entertainment firm that offers video-on-demand (VoD). Based on theories related to product sampling and information goods, we analyze the relationship between free previews and VoD purchases for series …

Towards A Science Of Security Games, Thanh Hong Nguyen, Debarun Kar, Matthew Brown, Arunesh Sinha, Albert Xin Jiang, Milind Tambe

Research Collection School Of Computing and Information Systems

Security is a critical concern around the world. In many domains from counter-terrorism to sustainability, limited security resources prevent full security coverage at all times; instead, these limited resources must be scheduled, while simultaneously taking into account different target priorities, the responses of the adversaries to the security posture and potential uncertainty over adversary types.Computational game theory can help design such security schedules. Indeed, casting the problem as a Bayesian Stackelberg game, we have developed new algorithms that are now deployed over multiple years in multiple applications for security scheduling. These applications are leading to real-world use-inspired research in the …

Visual Analysis Of Uncertainty In Trajectories, Lu Lu, Nan Cao, Siyuan Liu, Lionel Ni, Xiaoru Yuan, Huamin Qu

Research Collection School Of Computing and Information Systems

Mining trajectory datasets has many important applications. Real trajectory data often involve uncertainty due to inadequate sampling rates and measurement errors. For some trajectories, their precise positions cannot be recovered and the exact routes that vehicles traveled cannot be accurately reconstructed. In this paper, we investigate the uncertainty problem in trajectory data and present a visual analytics system to reveal, analyze, and solve the uncertainties associated with trajectory samples. We first propose two novel visual encoding schemes called the road map analyzer and the uncertainty lens for discovering road map errors and visually analyzing the uncertainty in trajectory data respectively. …

Optimizing Resolution And Uncertainty In Bathymetric Sonar Systems, Val E. Schmidt, Thomas C. Weber, Xavier Lurton

Center for Coastal and Ocean Mapping

Bathymetric sonar systems (whether multibeam or phase-differencing sidescan) contain an inherent trade-off between resolution and uncertainty. Systems are traditionally designed with a fixed spatial resolution, and the parameter settings are optimized to minimize the uncertainty in the soundings within that constraint. By fixing the spatial resolution of the system, current generation sonars operate sub-optimally when the SNR is high, producing soundings with lower resolution than is supportable by the data, and inefficiently when the SNR is low, producing high-uncertainty soundings of little value. Here we propose fixing the sounding measurement uncertainty instead, and optimizing the resolution of the system within …

Technology Investment Decision-Making Under Uncertainty: The Case Of Mobile Payment Systems, Robert J. Kauffman, Jun Liu, Dan Ma

Research Collection School Of Computing and Information Systems

The recent launch of Google Wallet has brought the issue of technology solutions in mobile payments (m-payments) to the forefront. In deciding whether and when to adopt m-payments, senior managers in banks are concerned about uncertainties regarding future market conditions, technology standards, and consumer and merchant responses, especially their willingness to adopt. This study applies economic theory and modeling for decision-making under uncertainty to bank investments in m-payment systems technology. We assess the projected benefits and costs of investment as a continuous-time stochastic process to determine optimal investment timing. We find that the value of waiting to adopt jumps when …

The Intelligence Game: Assessing Delphi Groups And Structured Question Formats, Bonnie Wintle, Steven Mascaro, Fiona Fidler, Marissa Mcbride, Mark Burgman, Louisa Flander, Geoff Saw, Charles Twardy, Aidan Lyon, Brian Manning

Australian Security and Intelligence Conference

In 2010, the US Intelligence Advanced Research Projects Activity (IARPA) announced a 4-year forecasting “tournament”. Five collaborative research teams are attempting to outperform a baseline opinion pool in predicting hundreds of geopolitical, economic and military events. We are contributing to one of these teams by eliciting forecasts from Delphi-style groups in the US and Australia. We elicit probabilities of outcomes for 3-5 monthly questions, such as: Will Australia formally transfer uranium to India by 1 June 2012? Participants submit probabilities in a 3-step interval format, view those of others in their group, share, rate and discuss information, and then make …

Temporal Data Mining Of Uncertain Water Reservoir Data, Abhinaya Mohan, Peter Revesz

CSE Conference and Workshop Papers

This paper describes the challenges of data mining uncertain water reservoir data based on past human operations in order to learn from them reservoir policies that can be automated for the future operation of the water reservoirs. Records of human operations of water reservoirs often contain uncertain data. For example, the recorded amounts of water released and retained in the water reservoirs are typically uncertain, i.e., they are bounded by some minimum and maximum values. Moreover, the time of release is also uncertain, i.e., typically only monthly or weekly amounts are recorded. To increase the effectiveness of data mining of …

Scalable Content Authentication In H.264/Svc Videos Using Perceptual Hashing Based On Dempster-Shafer Theory, Dengpan Ye, Zhuo Wei, Xuhua Ding, Robert H. Deng

Research Collection School Of Computing and Information Systems

The content authenticity of the multimedia delivery is important issue with rapid development and widely used of multimedia technology. Till now many authentication solutions had been proposed, such as cryptology and watermarking based methods. However, in latest heterogeneous network the video stream transmission has b een coded in scalable way such as H.264/SVC, there is still no good authentication solution. In this paper, we firstly summarized related works and p roposed a scalable content authentication scheme using a ratio of different energy (RDE) based perceptual hashing in Q/S dimension, which is used Dempster-Shafer theory and combined with the latest scalable …

How To Divide Students Into Groups So As To Optimize Learning: Towards A Solution To A Pedagogy-Related Optimization Problem, Olga Kosheleva, Vladik Kreinovich

Departmental Technical Reports (CS)

To enhance learning, it is desirable to also let students learn from each other, e.g., by working in groups. It is known that such groupwork can improve learning, but the effect strongly depends on how we divide students into groups. In this paper, based on a first approximation model of student interaction, we describe how to optimally divide students into groups so as to optimize the resulting learning. We hope that, by taking into account other aspects of student interaction, it will be possible to transform our solution into truly optimal practical recommendations.