arXiv subjects

Matt Turner

Publications and source records attributed to Matt Turner.

5 recordsLinked to original sources

arXiv2026

KnifeHunter: Structured Local Representation Learning for Fine-Grained Knife Image Retrieval in Law Enforcement

Knife-enabled violence presents a major public safety challenge, and law enforcement agencies require scalable tools for catalogue-level knife identification, intelligence analysis, and source attribution. Manual visual comparison is specialist, time-consuming, and difficult to scale under operational imaging conditions. We introduce KnifeHunter, an end-to-end forensic knife image retrieval system developed with UK law enforcement. The work contributes the KnifeHunter dataset, comprising 25,843 images across 543 knife classes from police evidence, retail catalogues, and border-force seizures, with structured metadata, Medium/Hard evaluation protocols, and large-scale distractor evaluation. We further propose CoRe-Net, a compact single-descriptor retrieval architecture that combines global context with spatially localised discriminative evidence. CoRe-Net introduces Structured Complementary Representation Learning (SCRL) to organise local evidence into complementary prototype-based representations, and Bi-Directional Reciprocal Fusion (BDRF) to integrate global and local evidence through residual projection and gated local-to-global injection. Using an EVA02-Base backbone and cosine-similarity retrieval, CoRe-Net achieves 88.0% mAP and 86.7% mP@10 on the Medium protocol, and 85.1% mAP and 83.8% mP@10 under distractor conditions. KnifeHunter was deployed by UK police forces during Operation Sceptre deployments from 2023 to 2025, achieving 99.2% mP@1 on field queries. These results demonstrate a practical and effective multimedia retrieval framework for fine-grained forensic knife matching in operational law-enforcement settings.

cs.CV↗

arXiv2025

Characterizing and Optimizing LLM Inference Workloads on CPU-GPU Coupled Architectures

Large language model (LLM)-based inference workloads increasingly dominate data center costs and resource utilization. Therefore, understanding the inference workload characteristics on evolving CPU-GPU coupled architectures is crucial for optimization. This paper presents an in-depth analysis of LLM inference behavior on loosely-coupled (PCIe A100/H100) and closely-coupled (GH200) systems. We analyze performance dynamics using fine-grained operator-to-kernel trace analysis, facilitated by our novel profiler SKIP and metrics like Total Kernel Launch and Queuing Time (TKLQT). Results show that closely-coupled (CC) GH200 significantly outperforms loosely-coupled (LC) systems at large batch sizes, achieving 1.9x-2.7x faster prefill latency for Llama 3.2-1B. However, our analysis also reveals that GH200 remains CPU-bound up to 4x larger batch sizes than LC systems. In this extended CPU-bound region, we identify the performance characteristics of the Grace CPU as a key factor contributing to higher inference latency at low batch sizes on GH200. We demonstrate that TKLQT accurately identifies this CPU/GPU-bound transition point. Based on this analysis, we further show that kernel fusion offers significant potential to mitigate GH200's low-batch latency bottleneck by reducing kernel launch overhead. This detailed kernel-level characterization provides critical insights for optimizing diverse CPU-GPU coupling strategies. This work is an initial effort, and we plan to explore other major AI/DL workloads that demand different degrees of CPU-GPU heterogeneous architectures.

cs.DC↗

arXiv2024

$G_2$-instantons on the ALC members of the $\mathbb{B}_7$ family

Using co-homogeneity one symmetries, we construct a two-parameter family of non-abelian $G_2$-instantons on every member of the asymptotically locally conical $\mathbb{B}_7$-family of $G_2$-metrics on $S^3 \times \mathbb{R}^4 $, and classify the resulting solutions. These solutions can be described as perturbations of a one-parameter family of abelian instantons, arising from the Killing vector-field generating the asymptotic circle fibre. Generically, these perturbations decay exponentially to the model, but we find a one-parameter family of instantons with polynomial decay. Moreover, we relate the two-parameter family to a lift of an explicit two-parameter family of anti-self-dual instantons on Taub-NUT $\mathbb{R}^4$, fibred over $S^3$ in an adiabatic limit.

math.DG↗

arXiv2023

$G_2$-instantons on the spinor bundle of the 3-sphere

We classify $G_2$-instantons admitting $SU(2)^3$-symmetries, and construct a new family of examples on the spinor bundle of the 3-sphere, equipped with the asymptotically conical, co-homogeneity one $G_2$-metric of Bryant-Salamon. We also show that, outside of the $SU(2)^3$-invariant examples, any other $G_2$-instanton on this metric with the same asymptotic behaviour must have obstructed deformations.

math.DG↗

arXiv2022

SU(2)^2xU(1)-invariant G_2-instantons on the AC limit of the C7 family

We construct SU(2)^2xU(1)-invariant G_2-instantons on the asymptotically conical limit of the C7 family of G_2-metrics. The construction uses a dynamical systems approach involving perturbations of an abelian solution and a solution on the G_2-cone. From this we obtain a 1-parameter family of invariant instantons with gauge group SU(2) and bounded curvature.

math.DG↗