Institution profile

University of Auckland

Academic institutionaustralasia · nz
Official website
Research library343linked papers
Opportunities0open roles
Selected work

Representative Papers

Accurate Forgetting for Heterogeneous Federated Continual Learning

Feb 20, 2025International Conference on Learning Representations

To address statistical bias and noise interference arising from client data/task heterogeneity—or even adversarial behavior—in federated continual learning (FCL), this paper introduces the “Accurate Forgetting” (AF) paradigm: proactively identifying and discarding unreliable feature representations induced by skewed distributions and noise prior to knowledge reuse. Methodologically, we propose the first probability-based credibility assessment framework built upon normalized flows, enabling quantifiable, knowledge-granular filtering. Further, we integrate generative replay with selective knowledge inheritance to dynamically enhance global model robustness within the federated architecture. Evaluated on multiple heterogeneous FCL benchmarks, AF achieves an average accuracy improvement of 12.3%, significantly boosting generalization and noise resilience. Our approach provides a novel, interpretable, and computationally tractable pathway for bias mitigation in FCL.

5 citationsRead paper

Metamorphic Testing of Large Language Models for Natural Language Processing

Sep 07, 2025IEEE International Conference on Software Maintenance and Evolution

This work addresses the challenge of automatically identifying erroneous behaviors of large language models (LLMs) in NLP tasks under data-scarce or zero-shot settings, where ground-truth labels are unavailable. We propose an oracle-free defect detection method based on metamorphic testing (MT), systematically constructing and validating 191 metamorphic relations covering semantic, syntactic, and task-specific logic—the most comprehensive MT study for LLMs to date. We select 36 representative relations and conduct 560,000 tests across three major LLM families, quantitatively measuring response consistency. Our experiments reveal, for the first time, systematic inconsistency in LLMs under diverse semantic-preserving transformations, empirically demonstrating MT’s efficacy in exposing robustness deficiencies. We further characterize MT’s applicability boundaries and inherent limitations. This work establishes a scalable, annotation-light paradigm for trustworthy LLM evaluation.

2 citationsRead paper

How similar are two elections?

Feb 01, 2025Journal of computer and system sciences (Print)

This study addresses the problem of measuring the similarity between two ordinal elections with identical numbers of candidates and voters, under the requirement that the measure be invariant under relabeling of both candidates and voters. To this end, the work proposes the first distance notion for elections that satisfies isomorphism invariance, defined by aligning candidate labels and finding an optimal bijection between voters that minimizes the total distance between matched preference orders. The authors show that election isomorphism can be decided in polynomial time, yet two natural variants of the isomorphism-invariant distance are both NP-complete and hard to approximate. Fixed-parameter tractable (FPT) algorithms are developed with respect to several natural parameters, including the number of voters and the number of distinct preference types. The technical approach integrates techniques from graph isomorphism testing, combinatorial optimization, and parameterized complexity analysis.

1 citationsRead paper
Recent publications

Latest Papers