While Multimodal Large Language Models (MLLMs) demonstrate remarkable capabilities across diverse domains, their application to specialized anomaly detection (AD) remains constrained by domain adaptation challenges. Existing Group Relative Policy Optimization (GRPO) based approaches suffer from two critical limitations: inadequate training data utilization when models produce uniform responses, and insufficient supervision over reasoning processes that encourage immediate binary decisions without deliberative analysis. We propose a comprehensive framework addressing these limitations through two synergistic innovations. First, we introduce a multi-stage deliberative reasoning process that guides models from region identification to focused examination, generating diverse response patterns essential for GRPO optimization while enabling structured supervision over analytical workflows. Second, we develop a fine-grained reward mechanism incorporating classification accuracy and localization supervision, transforming binary feedback into continuous signals that distinguish genuine analytical insight from spurious correctness. Comprehensive evaluation across multiple industrial datasets shows that our method achieves superior accuracy by enabling general-purpose MLLMs to acquire fine-grained visual discrimination for detecting subtle manufacturing defects.

Although video generation and editing models have advanced significantly, individual models remain restricted to specific tasks, often failing to meet diverse user needs. Effectively coordinating these models in pipelines can unlock a wide range of video generation and editing capabilities. However, manual orchestration is complex, time-consuming, and requires deep expertise in model performance and limitations. To address these challenges, we propose the Semantic Planning Agent (SPAgent), a novel system that automatically coordinates state-of-the-art open-source models to fulfill complex user intents. To equip SPAgent with robust orchestration capabilities, we introduce a three-step framework: 1) decoupled intent recognition to accurately parse multi-modal inputs; 2) principle-guided route planning to design effective execution chains; and 3) capability-based model selection to identify the optimal tools for each sub-task. To facilitate training, we curate a comprehensive multi-task generative video dataset. Furthermore, we enhance SPAgent with a video quality evaluation module, enabling it to autonomously assess and incorporate new models into its tool library without human intervention. Experimental results demonstrate that SPAgent effectively coordinates models to generate and edit high-quality videos, exhibiting superior versatility and adaptability across various tasks.

Diffusion Transformers (DiTs) have proven effective in generating high-quality videos but are hindered by high computational costs. Existing video diffusion sampling acceleration methods often rely on costly fine-tuning or exhibit limited generalization capabilities. We propose Asymmetric Reduction and Restoration (**AsymRnR**), **a training-free and model-agnostic method to accelerate video DiTs**. It builds on the observation that redundancies of feature tokens in DiTs vary significantly across different model blocks, denoising steps, and feature types. Our AsymRnR asymmetrically reduces redundant tokens in the attention operation, achieving acceleration with negligible degradation in output quality and, in some cases, even improving it. We also tailored a reduction schedule to distribute the reduction across components adaptively. To further accelerate this process, we introduce a matching cache for more efficient reduction. Backed by theoretical foundations and extensive experimental validation, AsymRnR integrates into state-of-the-art video DiTs and offers substantial speedup.

3D Visual Grounding (3DVG) aims to localize target objects within a 3D scene based on natural language queries. To alleviate the reliance on costly 3D training data, recent studies have explored zero-shot 3DVG by leveraging the extensive knowledge and powerful reasoning capabilities of pre-trained LLMs and VLMs. However, existing paradigms tend to emphasize either spatial (3D-based) or semantic (2D-based) understanding, limiting their effectiveness in complex real-world applications. In this work, we introduce SPAZER — a VLM-driven agent that combines both modalities in a progressive reasoning framework. It first holistically analyzes the scene and produces a 3D rendering from the optimal viewpoint. Based on this, anchor-guided candidate screening is conducted to perform a coarse-level localization of potential objects. Furthermore, leveraging retrieved relevant 2D camera images, 3D-2D joint decision-making is efficiently performed to determine the best-matching object. By bridging spatial and semantic reasoning neural streams, SPAZER achieves robust zero-shot grounding without training on 3D-labeled data. Extensive experiments on ScanRefer and Nr3D benchmarks demonstrate that SPAZER significantly outperforms previous state-of-the-art zero-shot methods, achieving notable gains of $\mathbf{9.0\}$% and $\mathbf{10.9\}$% in accuracy.

Video diffusion transformers have achieved remarkable progress in high-quality video generation, but remain computationally expensive due to the quadratic complexity of attention over high-dimensional video sequences. Recent acceleration methods enhance the efficiency by exploiting the local sparsity of attention scores; yet they often struggle with accelerating the long-range computation. To address this problem, we propose VORTA, an acceleration framework with two novel components: 1) a sparse attention mechanism that efficiently captures long-range dependencies, and 2) a routing strategy that adaptively replaces full 3D attention with specialized sparse attention variants. VORTA achieves an end-to-end speedup $1.76\times$ without loss of quality on VBench. Furthermore, it can seamlessly integrate with various other acceleration methods, such as model caching and step distillation, reaching up to speedup $14.41\times$ with negligible performance degradation. VORTA demonstrates its efficiency and enhances the practicality of video diffusion transformers in real-world settings. Codes and weights are available at https://github.com/wenhao728/VORTA.

The Ambient Internet of Things (AIoT), recently standardized by the 3rd Generation Partnership Project (3GPP), demands a low-power wide-area communication solution. Ambient Backscatter Communication (AmBC) is a promising approach that enables Backscatter Devices (BDs) to communicate by reusing existing signals without additional spectrum or energy cost. This paper investigates the feasibility of AmBC in the uplink of Long Term Evolution (LTE) cellular systems. We model the received BD signal, derive the optimal Maximum A Posteriori (MAP) detector, and propose simplified receiver architectures that achieve near-optimal performance with reduced complexity. A theoretical Bit Error Rate (BER) analysis is developed, including a Gaussian approximation, and the achievable coverage is evaluated. The results show that AmBC can sustain a BER on the order of 10−2at a reading distance of 1.5 m. Over-the-air experiments using commercial LTE uplink Sounding Reference Signals (SRS) further confirm the theoretical and simulation results. Overall, this work demonstrates that LTE cellular systems can natively support AmBC in the uplink, reinforcing its potential as a symbiotic radio paradigm for future low-power, spectrum-efficient AIoT connectivity.

Training deep object detectors demands expensive bounding box annotation. Active learning (AL) is a promising technique to alleviate the annotation burden. Performing AL at box-level for object detection, i.e., selecting the most informative boxes to label and supplementing the sparsely-labelled image with pseudo labels, has been shown to be more cost-effective than selecting and labelling the entire image. In box-level AL for object detection, we observe that models at early stage can only perform well on majority classes, making the pseudo labels severely class-imbalanced. We propose a class-balanced sampling strategy to select more objects from minority classes for labelling, so as to make the final training data, i.e., ground truth labels obtained by AL and pseudo labels, more class-balanced to train a better model. We also propose a task-aware soft pseudo labelling strategy to increase the accuracy of pseudo labels. We evaluate our method on public benchmarking datasets and show that our method achieves state-of-the-art performance.

The Long Term Evolution (LTE) is a promising can-didate to be used as a signal source in the Ambient Backscatter Communications (AmBC). LTE provides ubiquitous coverage, and LTE User Equipment (UE), such as mobile phone, can be used as the reader of a Backscatter Device (BD) signal with small firmware modifications. In this paper we measure a BD signal indoor coverage with commercial LTE base station as the illuminating ambient signal. The reader is synchronized to the LTE BS and estimates the BD signal from the LTE signal pilots. The BD is fully passive powered by a solar panel. The BD locations are distributed spatially along the corridor and for each BD location we measure the AmBC reader signal Bit Error Ratio (BER). The signal is uncoded. For each BD location the BER is measured from different directions with respect to the BD position. The measurement results show that along the corridor in most location, the observed BER is in order one percent. The results shed light on the feasibility of applications where the UE is in close proximity to the BD.

Existing approaches towards anomaly detection (AD) often rely on a substantial amount of anomaly-free data to train representation and density models. However, large anomaly-free datasets may not always be available before the inference stage; in which case an anomaly detection model must be trained with only a handful of normal samples, a.k.a. few-shot anomaly detection (FSAD). In this paper, we propose a novel methodology to address the challenge of FSAD which incorporates two important techniques. Firstly, we employ a model pre-trained on a large source dataset to initialize model weights. Secondly, to ameliorate the covariate shift between source and target domains, we adopt contrastive training to fine-tune on the few-shot target domain data. To learn suitable representations for the downstream AD task, we additionally incorporate cross-instance positive pairs to encourage a tight cluster of the normal samples, and negative pairs for better separation between normal and synthesized negative samples. We evaluate few-shot anomaly detection on 3 controlled AD tasks and 4 real-world AD tasks to demonstrate the effectiveness of the proposed method.

Deep learning (DL) algorithms have achieved unprecedented success in low-dose CT (LDCT) imaging and are expected to be a new generation of CT reconstruction technology. However, most DL-based denoising models often lack the ability to generalize to unseen dose data. Moreover, most simulation tools for LDCT typically operate on proprietary projection data, which is generally not accessible without an established collaboration with CT manufacturers. To alleviate these issues, in this work, we propose a dose-agnostic dual-task transfer network, termed DDT-Net, for simultaneous LDCT denoising and simulation. Concretely, the dual-task learning module is constructed to integrate the LDCT denoising and simulation tasks into a unified optimization framework by learning the joint distribution of LDCT and NDCT data. We approximate the joint distribution of continuous dose level data by training DDT-Net with discrete dose data, which can be generalized to denoising and simulation of unseen dose data. In particular, the mixed-dose training strategy adopted by DDT-Net can promote the denoising performance of lower-dose data. The paired dataset simulated by DDT-Net can be used for data augmentation to further restore the tissue texture of LDCT images. Experimental results on synthetic data and clinical data show that the proposed DDT-Net outperforms competing methods in terms of denoising and generalization performance at unseen dose data, and it also provides a simulation tool that can quickly simulate realistic LDCT images at arbitrary dose levels.

Ambient power enabled Internet of Things a.k.a. Ambient IoT is a novel concept connecting battery-free energy autonomous devices to the Internet. In this paper, we demonstrate the feasibility of using the channel estimator at the user equipment (UE) as a receiver for backscatter signals in LTE downlink. Our real-time demonstrator system consists of signal generator that mimics the operation of a base station (BS), a backscatter device (BD) modulating the incidence signal, and a software defined radio receiver that plays the role of the user equipment. The UE channel estimator and BD demodulation software is implemented with C++ and Python, respectively and made open source. Our results indicate that UE can demodulate the BD signal with bit error less than 10-3 when the downlink signal-to-noise ratio is larger than 10 dB when the pathloss between the BD and UE is less than 13 dB.

We implement several low-cost ambient backscatter communication (AmBC) receivers for the first time in practice. Multi-antenna AmBC receivers have been actively investigated theoretically. Practical implementations face a challenge as low-cost antennas are prone to suffer from wideband effect and/or hardware impairments and propagation environment is complicated. For a successful implementation, we propose a signal processing method which avoid the wideband signal but extracts the reference signal which is narrowband. We evaluate the method using the data measured in a complicated environment. Measurement results show that the proposed signal processing tackles the above issues and gives acceptable bit-error-rate (BER) performance of AmBC systems. The implementation of low-cost multi-antenna AmBC receivers paves the way for realizing ubiquitous green communication systems.

In this paper, the impact of transmit antenna separations at the base station on the downlink in a cellular OFDM system with per antenna rate control (PARC) is investigated. For the minimum mean square error (MMSE) with successive interference cancellation (SIC) receiver, the highly correlated interference resulting from closely spaced antennas is more easily suppressed than the independent interference resulting from widely separated antenna installations assuming the same number of interfering source(s). In the region where the interference from co-located source is dominating, the larger the transmit antenna separation, the better performances by the MMSE-SIC receiver. But in the region where interference from other cells is dominating, no gain is found when enlarging the transmit antenna separation, since it will only increase the number of uncorrelated interference sources. This means that for cell edge users it is beneficial with base stations with closely separated antenna installation in order to maximize the cancellation of inter-cell interference since the gain by multi-stream transmission for these users are small anyway.