Welcome to Research Commons

Research Commons is the University of Waikato's open access research repository, housing research publications and theses produced by the University's staff and students.

Communities in Research Commons

Select a community to browse its collections.

Now showing 1 - 2 of 2

Recent Submissions

Building adaptive knowledge bases for evolving continual learning models
(Springer Nature, 2025) Julian, Jack; Koh, Yun Sing; Bifet, Albert
Continual learning addresses catastrophic forgetting and knowledge transfer when learning from task streams. Dynamic architectures have introduced task-specific components like adapters layered over fixed pre-trained backbones. However, identifying the task of a new input remains a core challenge, leading to task-agnostic and dynamic detection methods. Existing approaches often overlook the reuse of previously learned adapters, missing opportunities for efficient forward and backwards transfer. We propose Continual Adapter-Based Learning (CABLE), a reinforcement learning framework that computes gradient similarity between new examples and past tasks. This similarity score drives a policy that assigns existing adapters when beneficial, rewarding improved performance and reducing reliance on newly initialised parameters. CABLE adopts a dynamic adapter routing strategy without assuming prior task labels. Evaluations on image classification and time series forecasting show that CABLE mitigates catastrophic forgetting and promotes efficient knowledge transfer across tasks.
The age of DDoScovery: An empirical comparison of industry and academic DDoS assessments
(ACM, 2024) Hiesgen, Raphael; Nawrocki, Marcin; Barcellos, Marinho; Kopp, Daniel; Hohlfeld, Oliver; Chan, Echo; Dobbins, Roland; Doerr, Christian; Rossow, Christian; Thomas, Daniel R.; Jonker, Mattijs; Mok, Ricky; Luo, Xiapu; Kristoff, John; Schmidt, Thomas C.; Wählisch, Matthias; Claffy, K. C.
Motivated by the impressive but diffuse scope of DDoS research and reporting, we undertake a multistakeholder (joint industry-academic) analysis to seek convergence across the best available macroscopic views of the relative trends in two dominant classes of attacks - direct-path attacks and reflection-amplification attacks. We first analyze 24 industry reports to extract trends and (in)consistencies across observations by commercial stakeholders in 2022. We then analyze ten data sets spanning industry and academic sources, across four years (2019-2023), to find and explain discrepancies based on data sources, vantage points, methods, and parameters. Our method includes a new approach: we share an aggregated list of DDoS targets with industry players who return the results of joining this list with their proprietary data sources to reveal gaps in visibility of the academic data sources. We use academic data sources to explore an industry-reported relative drop in spoofed reflection-amplification attacks in 2021-2022. Our study illustrates the value, but also the challenge, in independent validation of security-related properties of Internet infrastructure. Finally, we reflect on opportunities to facilitate greater common understanding of the DDoS landscape. We hope our results inform not only future academic and industry pursuits but also emerging policy efforts to reduce systemic Internet security vulnerabilities.
BusEnv: A multi-agent reinforcement learning environment and benchmark for urban public transportation
(ACM, 2026) da Silva e Silva, Wesley; Rios, Ricardo A.; da Costa Fonseca, Rafael; Ponnambalam, Sabarikirishwaran; Cassé, Léa; dos Santos Ferreira, Marcos Vinícius; Bifet, Albert; Rios, Tatiane N.
Reinforcement learning (RL) offers a powerful paradigm for managing complex, dynamic transportation systems where autonomous agents must adapt to uncertain and rapidly changing environments. We present BusEnv, a benchmark environment grounded in real-world data from the Salvador Urban Transportation Network, encompassing approximately 700,000 passengers, 2,000 vehicles, 400 lines, and 3,000 stops, collected between March 2024 and March 2025 at sub-minute resolution. BusEnv simulates realistic bus operations with stochastic passenger demand, route-specific travel times, and traffic-dependent variability, enabling controlled experimentation under partially observable, high-dimensional conditions. The reward function integrates multiple objectives, such as passenger service quality, operational efficiency, maintenance adherence, and sustainability, allowing the assessment of how different RL algorithms balance these competing factors. We evaluate nine baseline methods implemented in MARLlib, analyzing their convergence, robustness, and environmental impact when deployed under independent-learning conditions. Results show that PPO-based approaches achieve the highest stability and lowest energy waste, linking algorithmic robustness to sustainability performance. By combining data realism with reproducibility and extensibility, BusEnv establishes a foundation for systematic research on learning-based transport management and provides a scalable testbed for future studies on cooperative, sustainability-aware reinforcement learning.
The effects of passive and active warm-ups on vertical jump, stretch-shortening cycle and reactive strength index performance
(The University of Waikato, 2026-08-23) Li, Conglin; Rodrigues, Patrick
Passive warm-up, via hot-water immersion (HWI) for instance, may enhance explosive performance by increasing lower-limb temperature without imposing the metabolic demands associated with exercise. However, it remains unclear whether passive heating influences jump tasks with different concentric and stretch-shortening cycle (SSC) demands and whether its effects are comparable to those achieved following active warm-up (AWU). Hence, this thesis aimed to compare the acute effects of passive warm-up via HWI and AWU, on three different vertical jump performance tasks in physically active adults, including squat jump (SJ), countermovement jump (CMJ), and drop jump (DJ). Twenty participants (15 males and 5 females; age: 25.9 ± 4.5 years) completed the two warm-up conditions and a control condition (CON) via thermoneutral water immersion in a randomised crossover design. Passive warm-up involved 60 min of waist-level HWI at 42°C, whereas CON involved 60 min of immersion at 34°C. AWU comprised 15 min of cycling at a rating of perceived exertion of 14 (6-20 Borg scale), followed by dynamic lower-limb movements, 10 bodyweight squats, and 10 submaximal CMJs. The three vertical jump tests, SJ, CMJ, and DJ were assessed before and immediately after each condition on a force platform. The analysed outcomes were SJ and CMJ height and peak propulsive force (PPF), along with DJ height and reactive strength index (RSI). Change scores (post minus pre) were compared between conditions using analysis of variance (ANOVA) followed by Bonferroni-adjusted post hoc comparisons. Significant condition effects were observed for SJ height (p < .001), SJ PPF (p = .002), CMJ height (p < .001), DJ height (p = .006), and DJ RSI (p < .001). Bonferroni-adjusted comparisons showed that the positive changes observed following HWI differed significantly from the corresponding changes following AWU (SJ height, p < .001; SJ PPF, p = .020; CMJ height, p < .001; DJ height, p = .017; DJ RSI, p = .011) and CON (SJ height, p < .001; SJ PPF, p = .002; CMJ height, p < .001; DJ height, p = .015; DJ RSI, p < .001). CMJ PPF did not differ significantly among conditions (p = .108). These findings demonstrate that 60 min of waist-level HWI at 42°C can acutely improve concentric-dominant jump performance via SJ test and performance in jumps involving slow- and fast-SSC demands via CMJ and DJ, respectively. In addition, the passive warm-up via HWI enhanced RSI during DJ while AWU showed no difference to the CON condition. In conclusion, passive warm-up via 60 min of HWI at 42°C appears to be an effective strategy for acutely enhancing vertical jump performance across tasks characterised by concentric-dominant, slow-SSC, and fast-SSC demands.
Simulating the future, remembering the past: Effects of sensory and emotional focus on vividness and memory across the lifespan
(The University of Waikato, 2026) Kim, Joy; Devitt, Aleea
Remembering the past and simulating the future rely on overlapping constructive processes. Moreover, simulating a future event can influence what is later remembered. However, there are key gaps in our knowledge of this process when it comes to healthy ageing: older adults often report simulations as being more vivid than younger adults despite producing fewer episodic details, and they are less susceptible to simulation-induced memory biases. The present study aimed to address these gaps by manipulating the type of information prioritised during future simulation and examining influences on subjective vividness and subsequent memory across younger and older adults. Younger and older adults simulated future events while focusing on either sensory or emotional information, then read narratives describing the hypothetical outcomes of those events, before finally completing a recognition test for narrative details. Subjective vividness ratings were collected during simulation, and memory performance was assessed using discriminability (d′) and response bias (C). Despite clear differences in the type of details generated across sensory and emotion-focused conditions, neither vividness ratings nor memory accuracy (d′) varied as a function of focus condition. However, older adults demonstrated a more liberal response bias overall, and both age groups showed a more liberal bias following emotion-focused simulation. These findings suggest that the subjective experience of vividness is relatively independent of the specific type of information emphasised during simulation, indicating that vividness may reflect broader evaluative judgements about the simulation rather than the amount or type of episodic details the simulation contained. The absence of focus effects on memory accuracy further indicates that altering simulation content alone may be insufficient to bias later remembering. Instead, the influence of future simulation on memory may depend on whether the simulation meaningfully changes higher-level subjective states that persist into later encoding and retrieval.