Institution profile

University of Agder

Academic institutioneurope · no
Official website
Research library57linked papers
Opportunities0open roles
Selected work

Representative Papers

Scalable Temporal Anomaly Causality Discovery in Large Systems: Achieving Computational Efficiency with Binary Anomaly Flag Data

Dec 16, 2024arXiv.org

Causal discovery from binary alarm sequences in large-scale systems remains challenging due to the joint requirements of computational efficiency, sparse dependency modeling, and semantic capture of state transitions. Method: This paper proposes the first causal inference framework specifically designed for binary anomaly data. It introduces a sparse causal testing mechanism based on an improved Granger causality test, integrating flag-sequence feature encoding, adaptive graph-structure learning, and dynamic edge pruning. Contribution/Results: The framework explicitly models both the state-transition semantics and extreme sparsity inherent in binary data—novelty not addressed by prior work. By combining link compression with accuracy-aware pruning, it achieves scalable yet precise causal discovery. Evaluated on the CMS detector readout box system and IT monitoring datasets, it significantly reduces computational overhead while improving causal F1-score by a medium margin, thereby enabling effective real-time root-cause diagnosis.

1 citationsRead paper

TrimMoE A communication aware and adaptive depth framework for distributed edge inference

Aug 01, 2026

This work addresses the performance bottleneck caused by inter-server expert communication when deploying Mixture-of-Experts (MoE) large language models across distributed edge servers. The authors propose TrimMoE, a framework that jointly optimizes layer skipping, confidence-based early exiting, alternative execution, and server-expert selection to enable communication-aware, adaptive control of inference depth under a unified quality budget. Key innovations include lightweight per-layer exit heads, confidence-gated calibration, skip/early-exit-aware expert assignment, and proactive token migration prediction. Experiments on a 10-node heterogeneous platform demonstrate that TrimMoE reduces average latency by up to 62.8%, substantially decreases cross-server traffic and remote execution ratio, maintains stable throughput, and confines task quality degradation to within 2%.

0 citationsRead paper

HetRoute Heterogeneous and Cost-aware Collaborative Routing Framework for Distributed Edge MoE Inference

Aug 01, 2026

This work addresses the multi-dimensional cost optimization challenges in deploying Mixture-of-Experts (MoE) models across geographically distributed heterogeneous edge environments, where communication overhead, computational heterogeneity, GPU-CPU loading latency, queue backlogs, and quantization loss must be jointly managed. The authors propose HetRoute, a novel framework that introduces the first unified cost model to co-optimize expert placement, GPU-CPU residency, and quantization precision during offline deployment, while enabling holistic top-k expert routing during online inference to minimize bottleneck-layer costs. Integrating a cost-coupled deployment algorithm, beam-search-based routing, and queue- and quantization-aware scheduling, HetRoute achieves significant improvements on a 10-node heterogeneous edge testbed: it reduces average inference latency by 59.0%, P99 latency by 58.0%, inter-server traffic by 72.1%, and increases throughput by 2.13×, all while strictly bounding quality degradation within the prescribed budget.

0 citationsRead paper

Less is More: Modality-Decoupling for General AIGC Audio-Video Detection

Jul 28, 2026

This work addresses the limitations of existing audiovisual AIGC detection methods, which rely on the assumption of audio-visual consistency and often degrade in generalizable scenarios. To overcome this, we propose DAV-Det, the first modality-disentangled detection framework tailored for general-purpose AIGC content. Departing from conventional feature-level fusion, DAV-Det adopts decision-level fusion to independently model forgery traces in audio and video modalities. The visual branch leverages a three-level granularity representation—encompassing global, patch, and clip-level features—while the audio branch employs a gated dual-branch architecture operating in both time and frequency domains to capture anomalies. Evaluated on the IJCAI-ECAI 2026 DDL 2.0 Workshop challenge for general AIGC audiovisual detection, our method achieves state-of-the-art performance with a score of 0.8460, demonstrating superior robustness and generalization.

0 citationsRead paper
Recent publications

Latest Papers

TrimMoE A communication aware and adaptive depth framework for distributed edge inference

Aug 01, 2026

This work addresses the performance bottleneck caused by inter-server expert communication when deploying Mixture-of-Experts (MoE) large language models across distributed edge servers. The authors propose TrimMoE, a framework that jointly optimizes layer skipping, confidence-based early exiting, alternative execution, and server-expert selection to enable communication-aware, adaptive control of inference depth under a unified quality budget. Key innovations include lightweight per-layer exit heads, confidence-gated calibration, skip/early-exit-aware expert assignment, and proactive token migration prediction. Experiments on a 10-node heterogeneous platform demonstrate that TrimMoE reduces average latency by up to 62.8%, substantially decreases cross-server traffic and remote execution ratio, maintains stable throughput, and confines task quality degradation to within 2%.

0 citationsRead paper

HetRoute Heterogeneous and Cost-aware Collaborative Routing Framework for Distributed Edge MoE Inference

Aug 01, 2026

This work addresses the multi-dimensional cost optimization challenges in deploying Mixture-of-Experts (MoE) models across geographically distributed heterogeneous edge environments, where communication overhead, computational heterogeneity, GPU-CPU loading latency, queue backlogs, and quantization loss must be jointly managed. The authors propose HetRoute, a novel framework that introduces the first unified cost model to co-optimize expert placement, GPU-CPU residency, and quantization precision during offline deployment, while enabling holistic top-k expert routing during online inference to minimize bottleneck-layer costs. Integrating a cost-coupled deployment algorithm, beam-search-based routing, and queue- and quantization-aware scheduling, HetRoute achieves significant improvements on a 10-node heterogeneous edge testbed: it reduces average inference latency by 59.0%, P99 latency by 58.0%, inter-server traffic by 72.1%, and increases throughput by 2.13×, all while strictly bounding quality degradation within the prescribed budget.

0 citationsRead paper

Less is More: Modality-Decoupling for General AIGC Audio-Video Detection

Jul 28, 2026

This work addresses the limitations of existing audiovisual AIGC detection methods, which rely on the assumption of audio-visual consistency and often degrade in generalizable scenarios. To overcome this, we propose DAV-Det, the first modality-disentangled detection framework tailored for general-purpose AIGC content. Departing from conventional feature-level fusion, DAV-Det adopts decision-level fusion to independently model forgery traces in audio and video modalities. The visual branch leverages a three-level granularity representation—encompassing global, patch, and clip-level features—while the audio branch employs a gated dual-branch architecture operating in both time and frequency domains to capture anomalies. Evaluated on the IJCAI-ECAI 2026 DDL 2.0 Workshop challenge for general AIGC audiovisual detection, our method achieves state-of-the-art performance with a score of 0.8460, demonstrating superior robustness and generalization.

0 citationsRead paper

Pixel-Space Diffusion Transformers

Jul 20, 2026

This work addresses the limitations of latent diffusion models, which suffer from detail loss and misalignment between representation and generation objectives due to fixed visual encoders. To overcome these issues, the authors propose an end-to-end pixel-space diffusion Transformer framework that operates directly in the pixel domain without relying on VAE compression. The approach integrates a continuous generation mechanism within a unified multimodal Transformer architecture, sharing a common token space for both images and text. By carefully optimizing noise scheduling, loss weighting, and model scaling strategies, the method significantly enhances fine-grained detail fidelity in high-resolution image synthesis. This paradigm offers a promising direction toward building integrated multimodal vision foundation models capable of both generative and perceptual tasks.

0 citationsRead paper