A5087703163- Parallel Computing and Optimization Techniques
- Advanced Malware Detection Techniques
- Software Testing and Debugging Techniques
- Interactive and Immersive Displays
- Security and Verification in Computing
- Software-Defined Networks and 5G
- Scientific Computing and Data Management
- Emotion and Mood Recognition
- Topic Modeling
- IoT and Edge/Fog Computing
- Software System Performance and Reliability
- Tactile and Sensory Interactions
- Multimodal Machine Learning Applications
- Virtual Reality Applications and Impacts
- Digital Platforms and Economics
- Green IT and Sustainability
- Software Engineering Research
- Open Source Software Innovations
- Experimental Learning in Engineering
- Innovative Teaching and Learning Methods
- Advanced Data Storage Technologies
- Advanced Database Systems and Queries
- Advancements in Semiconductor Devices and Circuit Design
- Natural Language Processing Techniques
- Stroke Rehabilitation and Recovery
University of Auckland
2018-2025
Victoria University of Wellington
2016-2017
Morgan Stanley (United States)
1995
Abstract Retrieval Augment Generation (RAG) is a recent advancement in Open-Domain Question Answering (ODQA). RAG has only been trained and explored with Wikipedia-based external knowledge base not optimized for use other specialized domains such as healthcare news. In this paper, we evaluate the impact of joint training retriever generator components task domain adaptation ODQA. We propose RAG-end2end, an extension to that can adapt domain-specific by updating all during training. addition,...
Extracting emotions from physiological signals has become popular over the past decade. Recent advancements in wearable smart devices have enabled capturing continuously and unobtrusively. However, signal readings different wearables are lossy due to user activities, making it difficult develop robust models for emotion recognition. Also, limited availability of data labels is an inherent challenge developing machine learning techniques classification. This paper presents a novel...
Researchers have used machine learning approaches to identify motion sickness in VR experience. These would certainly benefit from an accurately labeled, real-world, diverse dataset that enables the development of generalizable ML models. We introduce 'VR.net', a comprising 165-hour gameplay videos 100 real-world games spanning ten genres, evaluated by 500 participants. VR.net assigns 24 sickness-related labels for each video frame, such as camera/object movement, depth field, and flow....
WebAssembly is a new-generation low-level byte-code format and gaining wide adoption in browser-centric applications. Nevertheless, originally designed as general approach for running binaries on any runtime environments more than the web. This paper presents Wasmachine, an OS aiming to efficiently securely execute applications IoT Fog devices with constrained resources. Wasmachine achieves efficient execution conventional OSs by compiling ahead of time native binary executing it kernel mode...
Distributed controller architectures have been proposed for Software-Defined Networking (SDN) to ensure scalability and reliability. One major drawback of the existing is uneven load distribution among controllers stemming from static binding between switches. To address this issue, several studies introduce dynamic by adopting some switch migration mechanisms that re-associate switches overloaded underutilized controllers. However, process adds a considerable amount complexity system may...
Smart devices are increasingly shrinking in size, which results new challenges for user-mobile interaction through minuscule touchscreens. Existing works to explore alternative technologies mainly rely on external degrade portability. In this paper, we propose UbiTouch, a novel system that extends smartphones with virtual touchpads desktops using built-in smartphone sensors. It senses user's finger movement proximity and ambient light sensor whose raw sensory data from underlying hardware...
Controller placement problem (CPP) is a key issue for Software-Defined Networking (SDN) with distributed controller architectures. This aims to determine suitable number of controllers deployed in important locations so as optimize the overall network performance. In comparison communication delay, existing literature on CPP assumes that influence workload distribution performance negligible. this paper, we tackle simultaneously considers control plane utilization, and distribution. Due...
WebAssembly is gaining wide adoption in browser-centric applications thanks to its high portability and security. More recently, researchers start exploring the potential of non-web environments, particularly IoT Edge computing. A key challenge performance gap with native code: compiled run slower by an average 45% as reported previous benchmark. The main causes slowdown are two-folded. Firstly, conventional runtimes translate instructions machine code using just-in-time compilers, which do...
Emotional Self-Awareness (ESA) plays a vital role in physical and mental well-being. Recent advancements artificial intelligence technologies have shown promising emotion recognition results, opening new opportunities to build systems support ESA. However, little research has been done understand users' perspectives on artificial-intelligence-based systems. We introduce Troi, an automatic mobile app using wearable signals. With we ran multi-day user study with 12 users preference parameters,...
Corneal Surface Reflections, or reflections on our eye-surface, have been shown as a valid and more socially acceptable source of information for passive lifelogging applications by prior work. However, automatic analysis corneal surface from single RGB camera to support is not extensively investigated in To address this, we developed synthetic self-supervised learning-based two-stage pipeline deep learning models detect objects these reflections. Our prototype only consists looking into the...
Despite the increasing popularity of VR games, one factor hindering industry's rapid growth is motion sickness experienced by users. Symptoms such as fatigue and nausea severely hamper user experience. Machine Learning methods could be used to automatically detect in experiences, but generating extensive labeled dataset needed a challenging task. It needs either very time consuming manual labeling human experts or modification proprietary application source codes for label capturing. To...
Competition in the semiconductor industry is forcing manufacturers to continuously improve capability of their equipment. The analysis real-time sensor data from manufacturing equipment presents opportunity reduce cost ownership Previous work by authors showed that time series filtering combination with multivariate techniques can be utilized perform statistical process control, and thereby generate alarms case malfunction. A more robust version this fault detection algorithm presented....
Many web applications have adopted WebAssembly thanks to its near-native performance and high portability. However, as a application grows complex, it starts suffering from bloated file size extensive startup time. This leads poor usability, especially on low-end devices. Existing works attempt address this issue by refactoring the codebase into several smaller shared libraries dynamically linking them demand. approach requires source code access human effort, which are not always feasible....
Mobile click fraud is a type of attack where an adversary deceptively generates events on mobile applications in pursuit revenue. Conventionally, the carried out by automating massive number physical devices. However, purchasing devices incur substantial costs. A cheaper alternative to emulators. existing emulators are inefficient and vastly blocked due their immense resource demand defective device signatures. In this paper, we propose Fraus <sup...
Open-source hardware such as RISC-V has been gaining substantial momentum. Recently, they have begun to embrace Google's Android operating system leverage its software ecosystem. Despite the encouraging progress, a challenging issue arises: majority of applications are written in native languages and need be recompiled target new platforms. Unfortunately, this recompilation process is not scalable because explosion To address issue, we present WasmAndroid, high-performance cross-platform...
Despite extensive security research on various Android components, such as kernel or runtime, little attention has been paid to the proprietary vendor blobs within firmware. In this paper, we conduct a large-scale empirical study understand update patterns and assess implications of blobs. We specifically focus GPU because they are loaded into every process for displaying graphics user interfaces can affect entire system's security. examine over 13,000 firmware releases between January 2018...
The text processing tool LATEX has prevailed as a standard in many fields of exact sciences; it is evident that likely to be here stay. From perspective, important explore what are the best possible ways support author efficiently editing documents. There have been several approaches provide graphical for LATEX. We argue true WYSIWYG (What You See Is What Get) approach justified requirement future systems and we present first cloud-based editor. This allows edit document its print form...
Researchers have used machine learning approaches to identify motion sickness in VR experience. These demand an accurately-labeled, real-world, and diverse dataset for high accuracy generalizability. As a starting point address this need, we introduce `VR.net', offering approximately 12-hour gameplay videos from ten real-world games 10 genres. For each video frame, rich set of sickness-related labels, such as camera/object movement, depth field, flow, are accurately assigned. Building is...
Technology integration in educational settings has led to the development of novel sensor-based tools that enable students measure and interact with their environment. Although reports from using such can be positive, evaluations are often conducted under controlled conditions short timeframes. There is a need for longitudinal data collected realistic classroom settings. However, sustained authentic use requires technology platforms seen by teachers as both easy value. We describe our...