Deep research agents have attracted growing attention for their potential to orchestrate multi-stage research workflows, spanning literature synthesis, methodological design, and empirical verification. Despite these strides, evaluating their research capability faithfully is rather challenging due to the difficulty of collecting frontier research questions that genuinely capture researchers’ attention and intellectual curiosity. To address this gap, we introduce DeepResearch Arena, a benchmark grounded in academic seminars that capture rich expert discourse and interaction, better reflecting real-world research environments and reducing the risk of data leakage. To automatically construct DeepResearch Arena, we propose a Multi-Agent Hierarchical Task Generation (MAHTG) system that extracts research-worthy inspirations from seminar transcripts. The MAHTG system further translates research-worthy inspirations into high-quality research tasks, ensuring the traceability of research task formulation while filtering noise. With the MAHTG system, we curate DeepResearch Arena with over 10,000 high-quality research tasks from over 200 academic seminars, spanning 12 disciplines, such as literature, history, and science. Our extensive evaluation shows that DeepResearch Arena presents substantial challenges for current state-of-the-art agents, with clear performance gaps observed across different models.

High-dimensional data present substantial challenges for distance-based classifiers, as the presence of many irrelevant or noisy features can reduce the informativeness of distance measures. Despite the availability of existing feature selection methods for such classifiers, the problem remains difficult in high-dimensional settings where noise can obscure class structure. In this paper, we propose a novel feature selection method grounded in class separability. To mitigate the effects of high dimensionality, the proposed approach evaluates each feature’s contribution to class separability by calculating separation scores across multiple randomly generated lower-dimensional feature subsets. We assess the proposed method using extensive simulation studies and 24 high-dimensional microarray datasets. The simulation results show that the method consistently identifies informative features while eliminating noise, thereby maintaining stable classification performance even when most features are non-informative. In benchmark comparisons with existing feature selection techniques, the proposed approach achieves superior predictive performance on most datasets while reducing dimensionality to, on average, 2% of the original feature set. A case study on prostate cancer data further demonstrates that the selected feature subset is enriched with biologically relevant signals. These findings suggest that the proposed method provides an effective and interpretable feature selection framework for distance-based classification of high-dimensional data.

One of the roadblocks for instance segmentation today is heavy computational overhead and model parameters. Previous methods based on Polar Representation made the initial mark to address this challenge by formulating instance segmentation as polygon detection, but failed to align with mainstream methods in performance. In this paper, we highlight that Representation Errors, arising from the limited capacity of polygons to capture boundary details, have long been overlooked, which results in severe performance degradation. Observing that optimal starting point selection effectively alleviates this issue, we propose an Adaptive Polygonal Sample Decision strategy to dynamically capture the positional variation of representation errors across samples. Additionally, we design a Union-aligned Rasterization Module to incorporate these errors into polygonal assessment, further advancing the proposed strategy. With these components, our framework PolarNeXt achieves a remarkable performance boost of over 4.8% AP compared to other polar-based methods. PolarNeXt is markedly more lightweight and efficient than state-of-the-art instance segmentation methods, while achieving comparable segmentation accuracy. We expect this work will open up a new direction for instance segmentation in high-resolution images and resource-limited scenarios. Codes can be found at https://github.com/Sun15194/PolarNeXt.

Device recognition is vital for security in wireless communication systems, particularly for applications like access control. Radio Frequency Fingerprint Identification (RFFI) offers a non-cryptographic solution by exploiting hardware-induced signal distortions. This paper proposes HyDRA, a Hybrid Dual-mode RF Architecture that integrates an optimized Variational Mode Decomposition (VMD) with a novel architecture based on the fusion of Convolutional Neural Networks (CNNs), Transformers, and Mamba components, designed to support both closed-set and open-set classification tasks. The optimized VMD enhances preprocessing efficiency and classification accuracy by fixing center frequencies and using closed-form solutions. HyDRA employs the Transformer Dynamic Sequence Encoder (TDSE) for global dependency modeling and the Mamba Linear Flow Encoder (MLFE) for linear-complexity processing, adapting to varying conditions. Evaluation on public datasets demonstrates state-of-the-art (SOTA) accuracy in closed-set scenarios and robust performance in our proposed open-set classification method, effectively identifying unauthorized devices. Deployed on NVIDIA Jetson Xavier NX, HyDRA achieves millisecond-level inference speed with low power consumption, providing a practical solution for real-time wireless authentication in real-world environments. The source code is published on https://github.com/Crazy-Bull/HyDRA.

The application of multispectral image based partial discharge detection offers a dependable solution for high-voltage substations. Captured visible light and ultraviolet (UV) images are denoised, transmitted and fused to enhance detection performance. However, existing approaches separately design the sensing-layer image denoising, communication-layer image transmission, and computing-layer image fusion, and the lack of unified cooperation hinders the overall performance. To address this issue, it is crucial to integrate sensing, communication, and computing to improve detection accuracy and timeliness. In this paper, we formulate a timeliness and accuracy joint guarantee problem, which aims to minimize the weighted sum of peak age of information (AoI), false-positive detection ratio, and false-negative detection ratio by jointly optimizing sensing-layer filtering window size, communication-layer time division ratio, and computing layer wavelet decomposition level. We propose a multispectral integrated sensing, communication, and computing algorithm based on AoI and false-negative aware multi-experience replay cooperative learning to solve the problem. Simulation results demonstrate that the proposed algorithm outperforms existing methods in terms of peak AoI, false-positive detection ratio, false-negative detection ratio, and convergence speed.

Action recognition from video data forms a cornerstone with wide-ranging applications. Single-view action recognition faces limitations due to its reliance on a single viewpoint. In contrast, multi-view approaches capture complementary information from various viewpoints for improved accuracy. Recently, event cameras have emerged as innovative bio-inspired sensors, leading to advancements in event-based action recognition. However, existing works predominantly focus on single-view scenarios, leaving a gap in multi-view event data exploitation, particularly in challenges like information deficit and semantic misalignment. To bridge this gap, we introduceHyperMV, multi-view event-based action recognition framework. HyperMV converts discrete event data into frame-like representations and extracts view-related features using a shared convolutional network. By treating segments as vertices and constructing hyperedges using rule-based and KNN-based strategies, a multi-view hypergraph neural network that captures relationships across viewpoint and temporal features is established. The vertex attention hypergraph propagation is also introduced for enhanced feature fusion. To prompt research in this area, we present the largest multi-view event-based action dataset$\mathbf{THU}^{\mathbf{MV-EACT}}\mathbf{-50}$, comprising 50 actions from 6 viewpoints, which surpasses existing datasets by over tenfold. Experimental results show that HyperMV significantly outperforms baselines in both cross-subject and cross-view scenarios, and also exceeds the state-of-the-arts in frame-based multi-view action recognition.

Approximate Computing (AC) saves energy and improves performance by introducing approximation into computation in error-torrent applications. This work focuses on an AC strategy that accurately performs important computations and approximates others. In order to determine which calculations are unimportant, we propose a novel importance evaluation algorithm, in which the key idea is a two-step fault injection to extract the near-optimal set of unimportant flip-flops in the circuit. The proposed algorithm reduces the complexity of architecture exploration from an exponential order to a linear order with-out understanding the functionality and behavior of the target application program.

Biological systems have a great number of visual motion detection neurons, some of which can preferentially react to specific visual regions. Nevertheless, little work has been performed about how they can be used to develop neural network models for omnidirectional collision detection. Hereby, an artificial fly visual brain neural network with presynaptic and postsynaptic subnetworks, for the first time, is developed to detect the changes of visual motion in panoramic scenes. Herein, the presynaptic subnetwork, which originates from the preferential response characteristics of five fly visual neurons, responds to all the moving objects in the panoramic field; the postsynaptic network, which is based on the properties of the angle and height detection neurons in the fly’s brain system, collects the excitatory intensities of the visual neurons, and outputs the real‐time activities of the main object closest to the panoramic camera. Hereafter, a computational model is constructed to implement omnidirectional collision detection, relying upon the artificial visual brain neural network and three functional neurons. The theoretical analysis has verified that the collision detection model’s computational complexity depends mainly on the image input resolution. Three experimental conclusions can be clearly drawn: (i) the motion characteristics of the main object in the panoramic environment can be clearly exhibited in the neural network; (ii) the collision detection model can not only outperform the compared models but also successfully perform omnidirectional collision detection; and (iii) it spends 0.24 s or so to execute visual information processing per frame with the resolution of 120 × 80.

Recent years have witnessed remarkable achievements in video-based action recognition. Apart from traditional frame-based cameras, event cameras are bio-inspired vision sensors that only record pixel-wise brightness changes rather than the brightness value. However, little effort has been made in event-based action recognition, and large-scale public datasets are also nearly unavailable. In this paper, we propose an event-based action recognition framework calledEV-ACT. The Learnable Multi-Fused Representation (LMFR) is first proposed to integrate multiple event information in a learnable manner. The LMFR with dual temporal granularity is fed into the event-based slow-fast network for the fusion of appearance and motion features. A spatial-temporal attention mechanism is introduced to further enhance the learning capability of action recognition. To prompt research in this direction, we have collected the largest event-based action recognition benchmark namedTHUE-ACT-50and the accompanyingTHUE-ACT-50-CHLdataset under challenging environments, including a total of over 12,830 recordings from 50 action categories, which is over 4 times the size of the previous largest dataset. Experimental results show that our proposed framework could achieve improvements of over 14.5%, 7.6%, 11.2%, and 7.4% compared to previous works on four benchmarks. We have also deployed our proposed EV-ACT framework on a mobile platform to validate its practicality and efficiency.

Person re- identification (Re-ID) has become a hot research topic due to its widespread applications. Conducting person Re-ID in video sequences is a practical requirement, in which the crucial challenge is how to pursue a robust video representation based on spatial and temporal features. However, most of the previous methods only consider how to integrate part-level features in the spatio-temporal range, while how to model and generate the part-correlations is little exploited. In this paper, we propose a skeleton-based dynamic hypergraph framework, namely Skeletal Temporal Dynamic Hypergraph Neural Network (ST-DHGNN) for person Re-ID, which resorts to modeling the high-order correlations among various body parts based on a time series of skeletal information. Specifically, multi-shape and multi-scale patches are heuristically cropped from feature maps, constituting spatial representations in different frames. A joint-centered hypergraph and a bone-centered hypergraph are constructed in parallel from multiple body parts (i.e., head, trunk, and legs) with spatio-temporal multi-granularity in the entire video sequence, in which the graph vertices representing regional features and hyperedges denoting relationships. Dynamic hypergraph propagation containing the re- planning module and the hyperedge elimination module is proposed to better integrate features among vertices. Feature aggregation and attention mechanisms are also adopted to obtain a better video representation for person Re-ID. Experiments show that the proposed method performs significantly better than the state-of-the-art on three video-based person Re-ID datasets, including iLIDS-VID, PRID-2011, and MARS.