Congestion control schemes based on information, such as queue size and traffic load, have become increasingly important, especially in data center networks, where the applications have stringent bandwidth and latency requirements. However, some congestion control schemes employ In-band Network Telemetry (INT), which introduces nontrivial bandwidth overhead. In this paper, we propose an efficient congestion control scheme: Precise ECN Encoding (PEE). PEE refactors the Explicit Congestion Notification (ECN) marking logic and proposes a novel ECN-based multi-packet joint encoding/decoding mechanism to enable more precise congestion perception. Compared with the state-of-the-art schemes, PEE achieves precise congestion control without introducing extra bandwidth overhead, achieving high throughput and low latency simultaneously. Comprehensive experimental results show that, compared to TIMELY, DCQCN, and HPCC, PEE reduces the average Flow Completion Time (FCT) by 102.1%, 26.8%, and 16.6% respectively under 80% network load.

Video anomaly detection (VAD) is vital for public safety, yet current approaches struggle with limited generalization, low interpretability, and high resource demands. To address these challenges, we propose HoloTrace, an edge-cloud collaborative VAD system that integrates large language models (LLMs) to construct and update a novel bidirectional causal knowledge graph. At the edge, HoloTrace leverages LLM-based cross-modal understanding and employs Hidden Markov Model (HMM) for bidirectional event reasoning, obtaining anomaly boundaries with low computational overhead. On the cloud side, LLMs are leveraged to dynamically update the Bi-CKG graph with key frames sent from the edge, in order to update causal relationships between events. Additionally, we introduce SVAD, a new large-scale VAD dataset comprising 632 real-world surveillance videos across 10 anomaly types and diverse scenes, with manually labeled frame-level annotations. Experimental results demonstrate that HoloTrace not only achieves the highest accuracy but also enhances interpretability and efficiency, paving the way for more generalizable and explainable video anomaly detection systems.

The success of computer vision tasks is mainly attributed to the architectural design of neural networks. This highlights the need to automatically design high-performance architectures via Neural Architecture Search (NAS). To accelerate the search process, training-free NAS is proposed, which aims to search high-performance architectures at initialization via zero-cost proxies (ZCPs). However, existing zero-cost proxies heavily rely on manual design, which is often labor-intensive and requires extensive expert knowledge. In addition, these crafted proxies often suffer from poor correlation with final model performance and high computational complexity, severely limiting NAS efficiency in real-world applications. To address those issues, this paper proposes a novel Large Language Models (LLMs)-driven $\underline{A}$utomatic $\underline{P}$roxy $\underline{D}$iscovery ($\textbf{APD}$) framework, which revolutionizes the design paradigm of ZCPs by leveraging LLMs to automatically discover optimal ZCPs for Training-Free NAS. Moreover, we utilize actor-critic based reinforcement learning to optimize prompts, enabling to generate better ZCPs in the next generation. We conduct extensive experiments on mainstream NAS benchmarks, demonstrating APD excels in both performance and efficiency. Besides, we firmly believe that our APD will dramatically benefit the deep learning community through providing novel paradigm of design algorithms via LLMs.

This paper investigates beamforming in a multiuser multiple-input multiple-output (MIMO) communication system assisted by a new type of circle track antenna (CTA) that can move along a circular trajectory. With this CTA structure, we formulate a sum rate maximization problem by jointly optimizing the beamforming and rotation angles of the CTAs, and transform it into an equivalent weighted minimum mean square error (WMMSE) minimization problem for efficient solution. To solve this highly non-convex problem, we use alternating optimization strategy to decompose the WMMSE problem into two sub-problems corresponding to the beamforming and the antenna rotation optimization, respectively. Specifically, in rotation angle optimization, we perform variable substitution to transform the optimization variables into ones with a constant modulus constraint and solve the problem using the manifold optimization (MO) method. Simulation results demonstrate that the proposed algorithm significantly outperforms the benchmark schemes.

The video analytics system utilizes deep learning models (DNN) to perform inference on the videos captured by cameras. Continuous learning algorithms are used to address the data drift problem in video analytics systems. However, uploading images from deployment environments and processing on the cloud carry the risk of privacy leakage. In this paper, we have designed a system called CL-Shield to protect user’s privacy. First, we review the causes of privacy leakage in a continuous learning system and propose the objective of full privacy protection. Second, we design an online training mechanism based on a scene library to avoid direct uploading of user’s frames to the cloud server. Lastly, we design a fast training set search algorithm based on a novel Ebv-List, which effectively improves the speed of model updates. We collect various real-world scenario data to build our scene library and validate our system on a dataset of over 10 hours. The experiments demonstrate that our privacy-aware continuous learning system achieves an F1-score of over 92% compared to the conventional systems without protecting privacy and has long-term stability in analytic F1-score.

Deploying deep neural network (DNN) models on resource-constrained edge devices for real-time video analytics poses significant challenges due to the high resource demands of these models. Current edge-based video analytics approaches often overlook optimizing deep learning models and GPU resource allocations in multi-tenant scenarios. In this paper, we present JSAS-MTMGS, a collaborative video analytics system employing three innovative design strategies. First, we propose a novel video configuration optimization space based on a joint DNN model sharing and splitting scheme to balance computational loads for collaborative processing. This approach reduces network transmission data volume and alleviates resource contention. Second, we design a GPU resource allocation scheme that combines GPU batching with spatial sharing to optimize GPU utilization and increase system throughput, all without relying on costly offline latency collection. Finally, we define the configuration optimization problem alongside GPU allocation as a convex problem and apply convex optimization to make scheduling decisions dynamically. Our experiments demonstrate that JSAS-MTMGS has the best service quality among all compared algorithms.

Real-time video analytics on edge devices has gained increasing attention across a wide range of business areas. However, edge devices usually have limited computing resources. Consequently, conventional approaches to video analytics either deploy simplified models on the edge (resulting in low accuracy) or transmit video content to the cloud (resulting in high latency and network overheads) to enable deep learning inference (e.g., object detection). In this paper, we introduce ParaLoupe, a novel real-time video analytics system that parallelizes deep learning inference in the edge cluster with task-oriented mini models. These mini models do not attain State-of-the-Art accuracy individually, but collectively can achieve much better accuracy-latency tradeoff than State-of-the-Art models. To achieve this, ParaLoupe crops multiple single-object patches from a given video frame. These single-object patches are then sent to multiple edge devices for parallel inference with specifically designed mini models. A patch-based task scheduling algorithm is further proposed to leverage the computing resources of the edge cluster to meet the service-level objectives. Our experimental results on real-world datasets show that ParaLoupe significantly outperforms baseline methods, achieving up to 14.1× inference speedup with accuracy on par with state-of-the-art models, or improving accuracy up to 45.1% under the same latency constraints.

Edge-cloud collaborative video analytics is transforming the way data is being handled, processed, and transmitted from the ever-growing number of surveillance cameras around the world. To avoid wasting limited bandwidth on unrelated content transmission, existing video analytics solutions usually perform temporal or spatial filtering to realize aggressive compression of irrelevant pixels. However, most of them work in a context-agnostic way while being oblivious to the circumstances where the video content is happening and the context-dependent characteristics under the hood. In this work, we propose VaBUS, a real-time video analytics system that leverages the rich contextual information of surveillance cameras to reduce bandwidth consumption for semantic compression. As a task-oriented communication system, VaBUS dynamically maintains the background image of the video on the edge with minimal system overhead and sends only highly confident Region of Interests (RoIs) to the cloud through adaptive weighting and encoding. With a lightweight experience-driven learning module, VaBUS is able to achieve high offline inference accuracy even when network congestion occurs. Experimental results show that VaBUS reduces bandwidth consumption by 25.0%-76.9% while achieving 90.7% accuracy for both the object detection and human keypoint detection tasks.

Unsupervised anomaly detection is a fundamental problem in various research areas and application domains, namely the discrimination of abnormal samples from normal samples where training data are only composed of one class (normal) while testing data contains both among which the majority are normal samples. However, previous works can not effectively fit the distribution of high dimensional data and suffers from low AUC scores which measures the classification performance of imbalanced data. To solve these problems, we propose an unsupervised anomaly detection model based on GAN, i.e., UAD-GAN. Specifically, we adopt transfer learning to extract visual features with pre-trained Inception-v3 model and use the discriminator to detect anomalies. UAD-GAN can fit the data distribution and detect anomalies efficiently. Extensive experiments show that UAD-GAN achieves state-of-the-art performance compared to other approaches.

Anomaly detection is an important area for both research and real-world applications. In the surface mounting technology (SMT) process, the defectives of solder paste printing need to be detected immediately or it may cause great effort for recycling and slow down the whole process. In this paper, we propose a novel model, MM-DNN, for anomaly detection with multi-modal data. We collect a multi-modal dataset from different sensors in the factory. Our method efficiently extracts both predictive features for classification and correlative features between multi-modal data to achieve a higher detection rate. As shown in the experiment, our method can further reduce 77% false alarm rate of the detection result in the factory while keeping 95% of real defectives be correctly detected.