arXiv subjects

Kevin Lee

Publications and source records attributed to Kevin Lee.

At least 19 recordsLinked to original sources

arXiv2026

Human-Aware Target Tracking and Navigation: Fusing Kinematic State Estimation with Structural Map Constraints

Autonomous mobile robots performing person-following tasks often suffer from temporary occlusions and sensor track loss in dynamic environments. This research presents an end-to-end autonomous navigation stack that addresses target occlusion through map-informed spatial reasoning. The proposed system features a multi-modal perception pipeline, fusing deep learning-based visual tracking with 2-dimensional LiDAR point clustering to maintain high-fidelity tracking of a tagged person. A continuous state estimator integrates this perception data with wheel odometry and IMU sensors for stable localization. When the active track is lost due to occlusion, the system activates a map-based recovery framework. Leveraging a predefined topological map, the system executes a graph-based search to propagate the target's last known trajectory along structurally defined walking lanes, adhering to left-hand regional conventions. By generating a discrete set of feasible future trajectories, the robot reasons about potential structural trajectory changes, such as continuing a heading or turning at an intersection. This map-informed prediction is fed directly to the local obstacle avoidance planner, enabling the robot to continue following its target safely and predictably until the person is visually reacquired. Real-world evaluations in dense multi-person environments demonstrate the system's robustness, achieving a 71.4\% target reacquisition success rate during major occlusion events lasting up to 7 seconds.

cs.RO↗

arXiv2026

Offline-to-Online Creative Optimization with Generative Models and Adaptive Testing

Ad creative optimization is increasingly constrained by evaluation rather than generation. Generative models can produce many plausible creatives, but reliable evaluation requires online experiments, in which only a limited slate can be tested. We study how to use data from historical A/B tests to generate and select the candidates in that slate. We developed and deployed a performance-driven offline-to-online workflow that guides creative generation with a predictive model as an inference-time critic. In the offline phase, we use a predictive model trained on historical experiments to rank and refine variants created by a generative model. A final test slate is then deployed in an online adaptive experiment. In a 50-arm field experiment, we found that the best creative generated with this method yielded 45.1% higher engagement than the best human-authored creative. Two additional experiments showed the same upper-tail pattern, with lifts of 46.7% and 36.2%. We found that despite the predictive model being too noisy to directly identify the best creative offline, it effectively guides the generative model toward creating strong candidates that can be efficiently evaluated in an adaptive experiment. The results suggest a design principle for creative optimization with generative models: use predictive models to guide generation of a slate to test, judge the slate by whether it contains high-performing candidates at a feasible test size, and use adaptive experiments to select among candidates while limiting traffic lost to weak arms.

cs.AI↗

arXiv2025

Always Keep Your Promises: A Model-Agnostic Attribution Algorithm for Neural Networks

Layer-wise Relevance Propagation (LRP) provides principled attribution for neural networks through conservation properties and foundations in Deep Taylor Decomposition. However, existing implementations operate at the module level, requiring architecture-specific propagation rules and model modifications. These limit the generality of target model and sustainability of implementations as architectures evolve. We introduce DynamicLRP, a model-agnostic LRP framework operating at the tensor operation level. By decomposing attribution to individual operations within computation graphs and introducing a novel mechanism for deferred activation resolution, named the Promise System, our approach achieves true architecture agnosticity while maintaining LRP's theoretical guarantees. This design operates independently of backpropagation machinery, requiring no model modification, enabling side-by-side execution with gradient backpropagation. Being based on computation graphs, this method is theoretically extensible to other deep learning libraries that support auto-differentiation. We demonstrate faithfulness matching or exceeding specialized implementations (1.77 vs 1.69 ABPC on VGG, equivalent performance on ViT, 93.70% and 95.06% top-1 attribution accuracy for explaining RoBERTa-large and Flan-T5-large answers on SQuADv2, respectively) while maintaining practical efficiency on models with 100M-1B parameters. We achieved 99.92% node coverage across 31,465 computation graph nodes from 15 diverse architectures, including state-space models (Mamba), audio transformers (Whisper), and multimodal systems (DePlot) without any model-specific code with rules for 47 fundamental operations implemented. Our operation-level decomposition and Promise System establish a sustainable, extensible foundation for LRP across evolving architectures. All code is available at https://github.com/keeinlev/dynamicLRP .

cs.LG↗

arXiv2025

Reasoning With a Star: A Heliophysics Dataset and Benchmark for Agentic Scientific Reasoning

Scientific reasoning through Large Language Models in heliophysics involves more than just recalling facts: it requires incorporating physical assumptions, maintaining consistent units, and providing clear scientific formats through coordinated approaches. To address these challenges, we present Reasoning With a Star, a newly contributed heliophysics dataset applicable to reasoning; we also provide an initial benchmarking approach. Our data are constructed from National Aeronautics and Space Administration & University Corporation for Atmospheric Research Living With a Star summer school problem sets and compiled into a readily consumable question-and-answer structure with question contexts, reasoning steps, expected answer type, ground-truth targets, format hints, and metadata. A programmatic grader checks the predictions using unit-aware numerical tolerance, symbolic equivalence, and schema validation. We benchmark a single-shot baseline and four multi-agent patterns, finding that decomposing workflows through systems engineering principles outperforms direct prompting on problems requiring deductive reasoning rather than pure inductive recall.

cs.AI↗

arXiv2025

Virtual Traffic Lights for Multi-Robot Navigation: Decentralized Planning with Centralized Conflict Resolution

We present a hybrid multi-robot coordination framework that combines decentralized path planning with centralized conflict resolution. In our approach, each robot autonomously plans its path and shares this information with a centralized node. The centralized system detects potential conflicts and allows only one of the conflicting robots to proceed at a time, instructing others to stop outside the conflicting area to avoid deadlocks. Unlike traditional centralized planning methods, our system does not dictate robot paths but instead provides stop commands, functioning as a virtual traffic light. In simulation experiments with multiple robots, our approach increased the success rate of robots reaching their goals while reducing deadlocks. Furthermore, we successfully validated the system in real-world experiments with two quadruped robots and separately with wheeled Duckiebots.

cs.RO↗

arXiv2025

SLDB: An End-To-End Heterogeneous System-on-Chip Benchmark Suite for LLM-Aided Design

Over the last few years, Large Language Models (LLMs) have emerged as a valuable tool for Electronic Design Automation (EDA). State-of-the-art research in LLM-aided design has demonstrated the ability of LLMs to generate syntactically correct RTL code, showcasing encouraging prospects for integrating AI into the hardware design process. A key enabler of these advancements is the availability of high-quality benchmarks to evaluate new approaches. However, existing datasets and benchmarks fall short of system-level design, as they focus primarily on component-level information and low-complexity designs. To address this gap, we introduce the System-Level Design Benchmark (SLDB), a dataset tailored for evaluating LLMs in system-level integration and configuration tasks. SLDB includes a curated benchmark suite of 10 baseline SoC designs, whose components can be combined into an exponential number of distinct tile-based SoCs through a synthetic library. The dataset provides full SoC configurations, accelerator integration code, communication parameters, and accelerator-aware system configurations, along with testing-application code, compatible with the ESP platform[1].

cs.AR↗

arXiv2025

Dangerous Questions in Astronomy Education

As astronomy enters an era defined by global telescope networks, petabyte-scale surveys, and powerful computational tools, the longstanding goals of astronomy education, particularly introductory ``ASTRO101'', but equally encompassing both higher and lower level courses, warrant fresh examination. In June 2024, the AstroEdUNC meeting at UNC--Chapel Hill convened 100 astronomers, education researchers, and practitioners to synthesise community perspectives on the purpose, content, and delivery of astronomy education. Beginning with historical vignettes, the meeting's deliberations were organised into six interrelated themes: (1) Context, highlighting astronomy's evolution from classical charting to multi-messenger discovery and its role as a connective thread across STEM and the humanities; (2) Content, exploring how curricula can balance essential concepts with authentic investigations and leverage open-source and AI-augmented resources; (3) Skills, arguing that astronomy should foreground scientific literacy, computational fluency, and communication through genuine data-driven inquiry; (4) Engagement, advocating for active-learning strategies, formative assessment, and culturally inclusive narratives; (5) Beyond the Classroom, emphasising scaffolding, universal-design practices, and K--12/community partnerships; and (6) Astronomy Education Research, outlining priority areas for assessing knowledge, attitudes, and long-term outcomes. We provide concrete recommendations for future astronomy education research development, underscoring the need for approaches to education that are authentic while meeting the learning and life goal needs of the students, a vibrant community of practice and robust researcher-practitioner partnerships to ensure that introductory astronomy is pertinent, applicable and inspiring to a broad student population.

physics.ed-ph↗

arXiv2025

The PanAf-FGBG Dataset: Understanding the Impact of Backgrounds in Wildlife Behaviour Recognition

Computer vision analysis of camera trap video footage is essential for wildlife conservation, as captured behaviours offer some of the earliest indicators of changes in population health. Recently, several high-impact animal behaviour datasets and methods have been introduced to encourage their use; however, the role of behaviour-correlated background information and its significant effect on out-of-distribution generalisation remain unexplored. In response, we present the PanAf-FGBG dataset, featuring 20 hours of wild chimpanzee behaviours, recorded at over 350 individual camera locations. Uniquely, it pairs every video with a chimpanzee (referred to as a foreground video) with a corresponding background video (with no chimpanzee) from the same camera location. We present two views of the dataset: one with overlapping camera locations and one with disjoint locations. This setup enables, for the first time, direct evaluation of in-distribution and out-of-distribution conditions, and for the impact of backgrounds on behaviour recognition models to be quantified. All clips come with rich behavioural annotations and metadata including unique camera IDs and detailed textual scene descriptions. Additionally, we establish several baselines and present a highly effective latent-space normalisation technique that boosts out-of-distribution performance by +5.42% mAP for convolutional and +3.75% mAP for transformer-based models. Finally, we provide an in-depth analysis on the role of backgrounds in out-of-distribution behaviour recognition, including the so far unexplored impact of background durations (i.e., the count of background frames within foreground videos).

cs.CV↗

arXiv2024

Iterative Encoding-Decoding VAEs Anomaly Detection in NOAA's DART Time Series: A Machine Learning Approach for Enhancing Data Integrity for NASA's GRACE-FO Verification and Validation

NOAA's Deep-ocean Assessment and Reporting of Tsunamis (DART) data are critical for NASA-JPL's tsunami detection, real-time operations, and oceanographic research. However, these time-series data often contain spikes, steps, and drifts that degrade data quality and obscure essential oceanographic features. To address these anomalies, the work introduces an Iterative Encoding-Decoding Variational Autoencoders (Iterative Encoding-Decoding VAEs) model to improve the quality of DART time series. Unlike traditional filtering and thresholding methods that risk distorting inherent signal characteristics, Iterative Encoding-Decoding VAEs progressively remove anomalies while preserving the data's latent structure. A hybrid thresholding approach further retains genuine oceanographic features near boundaries. Applied to complex DART datasets, this approach yields reconstructions that better maintain key oceanic properties compared to classical statistical techniques, offering improved robustness against spike removal and subtle step changes. The resulting high-quality data supports critical verification and validation efforts for the GRACE-FO mission at NASA-JPL, where accurate surface measurements are essential to modeling Earth's gravitational field and global water dynamics. Ultimately, this data processing method enhances tsunami detection and underpins future climate modeling with improved interpretability and reliability.

cs.LG↗

arXiv2024

Learned, Lagged, LLM-splained: LLM Responses to End User Security Questions

Answering end user security questions is challenging. While large language models (LLMs) like GPT, LLAMA, and Gemini are far from error-free, they have shown promise in answering a variety of questions outside of security. We studied LLM performance in the area of end user security by qualitatively evaluating 3 popular LLMs on 900 systematically collected end user security questions. While LLMs demonstrate broad generalist ``knowledge'' of end user security information, there are patterns of errors and limitations across LLMs consisting of stale and inaccurate answers, and indirect or unresponsive communication styles, all of which impacts the quality of information received. Based on these patterns, we suggest directions for model improvement and recommend user strategies for interacting with LLMs when seeking assistance with security.

cs.CR↗

arXiv2024

Towards Generalized On-Chip Communication for Programmable Accelerators in Heterogeneous Architectures

We present several enhancements to the open-source ESP platform to support flexible and efficient on-chip communication for programmable accelerators in heterogeneous SoCs. These enhancements include 1) a flexible point-to-point communication mechanism between accelerators, 2) a multicast NoC that supports data forwarding to multiple accelerators simultaneously, 3) accelerator synchronization leveraging the SoC's coherence protocol, 4) an accelerator interface that offers fine-grained control over the communication mode used, and 5) an example ISA extension to support our enhancements. Our solution adds negligible area to the SoC architecture and requires minimal changes to the accelerators themselves. We have validated most of these features in complex FPGA prototypes and plan to include them in the open-source release of ESP in the coming months.

cs.AR↗

arXiv2024

Shortchanged: Uncovering and Analyzing Intimate Partner Financial Abuse in Consumer Complaints

Digital financial services can introduce new digital-safety risks for users, particularly survivors of intimate partner financial abuse (IPFA). To offer improved support for such users, a comprehensive understanding of their support needs and the barriers they face to redress by financial institutions is essential. Drawing from a dataset of 2.7 million customer complaints, we implement a bespoke workflow that utilizes language-modeling techniques and expert human review to identify complaints describing IPFA. Our mixed-method analysis provides insight into the most common digital financial products involved in these attacks, and the barriers consumers report encountering when doing so. Our contributions are twofold; we offer the first human-labeled dataset for this overlooked harm and provide practical implications for technical practice, research, and design for better supporting and protecting survivors of IPFA.

cs.CY↗

arXiv2024

PanAf20K: A Large Video Dataset for Wild Ape Detection and Behaviour Recognition

We present the PanAf20K dataset, the largest and most diverse open-access annotated video dataset of great apes in their natural environment. It comprises more than 7 million frames across ~20,000 camera trap videos of chimpanzees and gorillas collected at 14 field sites in tropical Africa as part of the Pan African Programme: The Cultured Chimpanzee. The footage is accompanied by a rich set of annotations and benchmarks making it suitable for training and testing a variety of challenging and ecologically important computer vision tasks including ape detection and behaviour recognition. Furthering AI analysis of camera trap information is critical given the International Union for Conservation of Nature now lists all species in the great ape family as either Endangered or Critically Endangered. We hope the dataset can form a solid basis for engagement of the AI community to improve performance, efficiency, and result interpretation in order to support assessments of great ape presence, abundance, distribution, and behaviour and thereby aid conservation efforts.

cs.CV↗

arXiv2023

Persona-based Assessment of Software Engineering Student Research Projects: An Experience Report

Students enrolled in software engineering degrees are generally required to undertake a research project in their final year through which they demonstrate the ability to conduct research, communicate outcomes, and build in-depth expertise in an area. Assessment in these projects typically involves evaluating the product of their research via a thesis or a similar artifact. However, this misses a range of other factors that go into producing successful software engineers and researchers. Incorporating aspects such as process, attitudes, project complexity, and supervision support into the assessment can provide a more holistic evaluation of the performance likely to better align with the intended learning outcomes. In this paper, we present on our experience of adopting an innovative assessment approach to enhance learning outcomes and research performance in our software engineering research projects. Our approach adopted a task-oriented approach to portfolio assessment that incorporates student personas, frequent formative feedback, delayed summative grading, and standards-aligned outcomes-based assessment. We report upon our continuous improvement journey in adapting tasks and criteria to address the challenges of assessing student research projects. Our lessons learnt demonstrate the value of personas to guide the development of holistic rubrics, giving meaning to grades and focusing staff and student attention on attitudes and skills rather than a product only.

cs.SE↗

arXiv2022

Security policy audits: why and how

Information security isn't just about software and hardware -- it's at least as much about policies and processes. But the research community overwhelmingly focuses on the former over the latter, while gaping policy and process problems persist. In this experience paper, we describe a series of security policy audits that we conducted, exposing policy flaws affecting billions of users that can be -- and often are -- exploited by low-tech attackers who don't need to use any tools or exploit software vulnerabilities. The solutions, in turn, need to be policy-based. We advocate for the study of policies and processes, point out its intellectual and practical challenges, lay out our theory of change, and present a research agenda.

cs.CR↗

arXiv2022

Adversarial Attacks on Speech Recognition Systems for Mission-Critical Applications: A Survey

A Machine-Critical Application is a system that is fundamentally necessary to the success of specific and sensitive operations such as search and recovery, rescue, military, and emergency management actions. Recent advances in Machine Learning, Natural Language Processing, voice recognition, and speech processing technologies have naturally allowed the development and deployment of speech-based conversational interfaces to interact with various machine-critical applications. While these conversational interfaces have allowed users to give voice commands to carry out strategic and critical activities, their robustness to adversarial attacks remains uncertain and unclear. Indeed, Adversarial Artificial Intelligence (AI) which refers to a set of techniques that attempt to fool machine learning models with deceptive data, is a growing threat in the AI and machine learning research community, in particular for machine-critical applications. The most common reason of adversarial attacks is to cause a malfunction in a machine learning model. An adversarial attack might entail presenting a model with inaccurate or fabricated samples as it's training data, or introducing maliciously designed data to deceive an already trained model. While focusing on speech recognition for machine-critical applications, in this paper, we first review existing speech recognition techniques, then, we investigate the effectiveness of adversarial attacks and defenses against these systems, before outlining research challenges, defense recommendations, and future work. This paper is expected to serve researchers and practitioners as a reference to help them in understanding the challenges, position themselves and, ultimately, help them to improve existing models of speech recognition for mission-critical applications. Keywords: Mission-Critical Applications, Adversarial AI, Speech Recognition Systems.

cs.SD↗

arXiv2021

Dislocation and Indium Droplet Related Emission Inhomogeneities in InGaN LEDs

This report classifies emission inhomogeneities that manifest in InGaN quantum well blue light-emitting diodes grown by plasma-assisted molecular beam epitaxy on free-standing GaN substrates. By a combination of spatially resolved electroluminescence and cathodoluminescence measurements, atomic force microscopy, scanning electron microscopy and hot wet KOH etching, the identified inhomogeneities are found to fall in four categories. Labeled here as type I through IV, they are distinguishable by their size, density, energy, intensity, radiative and electronic characteristics and chemical etch pits which correlates them with dislocations. Type I exhibits a blueshift of about 120 meV for the InGaN quantum well emission attributed to a perturbation of the active region, which is related to indium droplets that form on the surface in the metal-rich InGaN growth condition. Specifically, we attribute the blueshift to a decreased growth rate of and indium incorporation in the InGaN quantum wells underneath the droplet which is postulated to be the result of reduced incorporated N species due to increased N$_{2}$ formation. The location of droplets are correlated with mixed type dislocations for type I defects. Types II through IV are due to screw dislocations, edge dislocations, and dislocation bunching, respectively, and form dark spots due to leakage current and nonradiative recombination.

cond-mat.mtrl-sci↗

arXiv2020

Adapting Security Warnings to Counter Online Disinformation

Disinformation is proliferating on the internet, and platforms are responding by attaching warnings to content. There is little evidence, however, that these warnings help users identify or avoid disinformation. In this work, we adapt methods and results from the information security warning literature in order to design and evaluate effective disinformation warnings. In an initial laboratory study, we used a simulated search task to examine contextual and interstitial disinformation warning designs. We found that users routinely ignore contextual warnings, but users notice interstitial warnings -- and respond by seeking information from alternative sources. We then conducted a follow-on crowdworker study with eight interstitial warning designs. We confirmed a significant impact on user information-seeking behavior, and we found that a warning's design could effectively inform users or convey a risk of harm. We also found, however, that neither user comprehension nor fear of harm moderated behavioral effects. Our work provides evidence that disinformation warnings can -- when designed well -- help users identify and avoid disinformation. We show a path forward for designing effective warnings, and we contribute repeatable methods for evaluating behavioral effects. We also surface a possible dilemma: disinformation warnings might be able to inform users and guide behavior, but the behavioral effects might result from user experience friction, not informed decision making.

cs.HC↗