Zhiqi Zhang

Tonghao Zhuang, Shanglong Hu, Yongsheng Luo et al.

We present a semi-supervised framework for joint segmentation and classification of fetal cardiac ultrasound images. Built upon the EchoCare multi-task backbone, our method integrates SAM-Med2D for boundary refinement and leverages DINOv3 to enhance pseudo-label quality. We introduce view-specific hard masking along with a two-stage optimization strategy: an EMA phase to consolidate segmentation capabilities, followed by a Classification Fine-Tuning phase that freezes segmentation parameters and resets the classification head to recover classification performance without compromising segmentation gains. Evaluated on the FETUS 2026 leaderboard, our method achieves a Dice Similarity Coefficient at 79.99%, Normalized Surface Distance at 61.62%, and F1-score at 41.20%, validating the effectiveness of our approach for prenatal congenital heart disease screening. Source code is publicly available at: https://github.com/2826056177/zcst_fetus2026.

Chenbin Pan, Zhiqi Zhang, Senem Velipasalar et al.

Capturing interaction of hands with objects is important to autonomously detect human actions from egocentric videos. In this work, we present a pyramid video transformer with a dynamic class token generator for egocentric action recognition. Different from previous video transformers, which use the same static embedding as the class token for diverse inputs, we propose a dynamic class token generator that produces a class token for each input video by analyzing the hand-object interaction and the related motion information. The dynamic class token can diffuse such information to the entire model by communicating with other informative tokens in the subsequent transformer layers. With the dynamic class token, dissimilarity between videos can be more prominent, which helps the model distinguish various inputs. In addition, traditional video transformers explore temporal features globally, which requires large amounts of computation. However, egocentric videos often have a large amount of background scene transition, which causes discontinuities across distant frames. In this case, blindly reducing the temporal sampling rate will risk losing crucial information. Hence, we also propose a pyramid architecture to hierarchically process the video from short-term high rate to long-term low rate. With the proposed architecture, we significantly reduce the computational cost as well as the memory requirement without sacrificing from the model performance. We perform comparisons with different baseline video transformers on the EPIC-KITCHENS-100 and EGTEA Gaze+ datasets. Both quantitative and qualitative results show that the proposed model can efficiently improve the performance for egocentric action recognition.

Zhiqi Zhang, Wen Lu, Jinshan Cao et al.

Land cover classification is a multi-class segmentation task to classify each pixel into a certain natural or man-made category of the earth surface, such as water, soil, natural vegetation, crops, and human infrastructure. Limited by hardware computational resources and memory capacity, most existing studies preprocessed original remote sensing images by down sampling or cropping them into small patches less than 512*512 pixels before sending them to a deep neural network. However, down sampling images incurs spatial detail loss, renders small segments hard to discriminate, and reverses the spatial resolution progress obtained by decades of years of efforts. Cropping images into small patches causes a loss of long-range context information, and restoring the predicted results to their original size brings extra latency. In response to the above weaknesses, we present an efficient lightweight semantic segmentation network termed MKANet. Aimed at the characteristics of top view high-resolution remote sensing imagery, MKANet utilizes sharing kernels to simultaneously and equally handle ground segments of inconsistent scales, and also employs parallel and shallow architecture to boost inference speed and friendly support image patches more than 10X larger. To enhance boundary and small segments discrimination, we also propose a method that captures category impurity areas, exploits boundary information and exerts an extra penalty on boundaries and small segment misjudgment. Both visual interpretations and quantitative metrics of extensive experiments demonstrate that MKANet acquires state-of-the-art accuracy on two land-cover classification datasets and infers 2X faster than other competitive lightweight networks. All these merits highlight the potential of MKANet in practical applications.

Zhiqi Zhang, Nitin Bansal, Changjiang Cai et al.

Synthetic datasets are often used to pretrain end-to-end optical flow networks, due to the lack of a large amount of labeled, real-scene data. But major drops in accuracy occur when moving from synthetic to real scenes. How do we better transfer the knowledge learned from synthetic to real domains? To this end, we propose CLIP-FLow, a semi-supervised iterative pseudo-labeling framework to transfer the pretraining knowledge to the target real domain. We leverage large-scale, unlabeled real data to facilitate transfer learning with the supervision of iteratively updated pseudo-ground truth labels, bridging the domain gap between the synthetic and the real. In addition, we propose a contrastive flow loss on reference features and the warped features by pseudo ground truth flows, to further boost the accurate matching and dampen the mismatching due to motion, occlusion, or noisy pseudo labels. We adopt RAFT as the backbone and obtain an F1-all error of 4.11%, i.e. a 19% error reduction from RAFT (5.10%) and ranking 2$^{nd}$ place at submission on the KITTI 2015 benchmark. Our framework can also be extended to other models, e.g. CRAFT, reducing the F1-all error from 4.79% to 4.66% on KITTI 2015 benchmark.

Yi Sun, Zhiqi Zhang, Xinhao Zhong et al.

Recent advances in flow matching models have significantly improved text-to-image generation quality, but also introduce growing safety risks due to the generation of harmful or undesirable content. Existing concept erasure methods are either inference-time interventions with limited effectiveness or rely on supervised fine-tuning (SFT), which requires precisely aligned data and struggles with scalability and multi-concept settings. In this paper, we propose \emph{FlowErase-RL}, the first GRPO-based framework for concept erasure in flow matching models. We reformulate concept erasure as a reward optimization problem and introduce a \textbf{dynamic dual-path reward mechanism} that jointly optimizes (i) a Concept Erasure (CE) reward to suppress target concepts and (ii) a Non-target Space (NS) reward to preserve generative fidelity. The two reward paths are adaptively balanced during training via a performance-driven switching strategy, enabling stable optimization without explicit supervision. Extensive experiments on nudity, object, and artistic style erasure demonstrate that our method achieves state-of-the-art erasure performance while maintaining strong image quality and semantic alignment. Moreover, it exhibits robust resistance to adversarial attacks and scales effectively to multi-concept scenarios. Our results establish a new paradigm for safe and controllable generation in flow matching models.

Shilei Luo, Zhiqi Zhang, Hengchen Dai et al.

AI agents powered by large language models are increasingly acting on behalf of humans in social and economic environments. Prior research has focused on their task performance and effects on human outcomes, but less is known about the relationship between agents and the specific individuals who deploy them. We ask whether agents systematically reflect the behavioral characteristics of their human owners, functioning as behavioral extensions rather than producing generic outputs. We study this question using 10,659 matched human-agent pairs from Moltbook, a social media platform where each autonomous agent is publicly linked to its owner's Twitter/X account. By comparing agents' posts on Moltbook with their owners' Twitter/X activity across features spanning topics, values, affect, and linguistic style, we find systematic transfer between agents and their specific owners. This transfer persists among agents without explicit configuration, and pairs that align on one behavioral dimension tend to align on others. These patterns are consistent with transfer emerging through accumulated interaction between owners (or owners' computer environments) and their agents in everyday use. We further show that agents with stronger behavioral transfer are more likely to disclose owner-related personal information in public discourse, suggesting that the same owner-specific context that drives behavioral transfer may also create privacy risk during ordinary use. Taken together, our results indicate that AI agents do not simply generate content, but reflect owner-related context in ways that can propagate human behavioral heterogeneity into digital environments, with implications for privacy, platform design, and the governance of agentic systems.

Zhiqi Zhang, Xinhao Zhong, Yi Sun et al.

Text-to-image diffusion models have demonstrated remarkable capabilities in generating high-quality images, yet their tendency to reproduce undesirable concepts, such as NSFW content, copyrighted styles, or specific objects, poses growing concerns for safe and controllable deployment. While existing concept erasure approaches primarily focus on DDPM-based diffusion models and rely on costly fine-tuning, the recent emergence of flow matching models introduces a fundamentally different generative paradigm for which prior methods are not directly applicable. In this paper, we propose Differential Vector Erasure (DVE), a training-free concept erasure method specifically designed for flow matching models. Our key insight is that semantic concepts are implicitly encoded in the directional structure of the velocity field governing the generative flow. Leveraging this observation, we construct a differential vector field that characterizes the directional discrepancy between a target concept and a carefully chosen anchor concept. During inference, DVE selectively removes concept-specific components by projecting the velocity field onto the differential direction, enabling precise concept suppression without affecting irrelevant semantics. Extensive experiments on FLUX demonstrate that DVE consistently outperforms existing baselines on a wide range of concept erasure tasks, including NSFW suppression, artistic style removal, and object erasure, while preserving image quality and diversity.

Xinhao Zhong, Yimin Zhou, Zhiqi Zhang et al.

The rapid progress of visual autoregressive (VAR) models has brought new opportunities for text-to-image generation, but also heightened safety concerns. Existing concept erasure techniques, primarily designed for diffusion models, fail to generalize to VARs due to their next-scale token prediction paradigm. In this paper, we first propose a novel VAR Erasure framework VARE that enables stable concept erasure in VAR models by leveraging auxiliary visual tokens to reduce fine-tuning intensity. Building upon this, we introduce S-VARE, a novel and effective concept erasure method designed for VAR, which incorporates a filtered cross entropy loss to precisely identify and minimally adjust unsafe visual tokens, along with a preservation loss to maintain semantic fidelity, addressing the issues such as language drift and reduced diversity introduce by naïve fine-tuning. Extensive experiments demonstrate that our approach achieves surgical concept erasure while preserving generation quality, thereby closing the safety gap in autoregressive text-to-image generation by earlier methods.

Haotong Cheng, Zhiqi Zhang, Hao Li et al.

Deep learning has substantially advanced the field of Single Image Super-Resolution (SISR). However, existing research has predominantly focused on raw performance gains, with little attention paid to quantifying the transferability of architectural components. In this paper, we introduce the concept of "Universality" and its associated definitions, which extend the traditional notion of "Generalization" to encompass the ease of transferability of modules. We then propose the Universality Assessment Equation (UAE), a metric that quantifies how readily a given module can be transplanted across models and reveals the combined influence of multiple existing metrics on transferability. Guided by the UAE results of standard residual blocks and other plug-and-play modules, we further design two optimized modules: the Cycle Residual Block (CRB) and the Depth-Wise Cycle Residual Block (DCRB). Through comprehensive experiments on natural-scene benchmarks, remote-sensing datasets, and other low-level tasks, we demonstrate that networks embedded with the proposed plug-and-play modules outperform several state-of-the-art methods, achieving a PSNR improvement of up to 0.83 dB or enabling a 71.3% reduction in parameters with negligible loss in reconstruction fidelity. Similar optimization approaches could be applied to a broader range of basic modules, offering a new paradigm for the design of plug-and-play modules.

Rui Song, Xingbing Chen, Zelong Liu et al.

One of the key problems in multi-label text classification is how to take advantage of the correlation among labels. However, it is very challenging to directly model the correlations among labels in a complex and unknown label space. In this paper, we propose a Label Mask multi-label text classification model (LM-MTC), which is inspired by the idea of cloze questions of language model. LM-MTC is able to capture implicit relationships among labels through the powerful ability of pre-train language models. On the basis, we assign a different token to each potential label, and randomly mask the token with a certain probability to build a label based Masked Language Model (MLM). We train the MTC and MLM together, further improving the generalization ability of the model. A large number of experiments on multiple datasets demonstrate the effectiveness of our method.

Tongwen Huang, Zhiqi Zhang, Junlin Zhang

Advertising and feed ranking are essential to many Internet companies such as Facebook and Sina Weibo. Among many real-world advertising and feed ranking systems, click through rate (CTR) prediction plays a central role. There are many proposed models in this field such as logistic regression, tree based models, factorization machine based models and deep learning based CTR models. However, many current works calculate the feature interactions in a simple way such as Hadamard product and inner product and they care less about the importance of features. In this paper, a new model named FiBiNET as an abbreviation for Feature Importance and Bilinear feature Interaction NETwork is proposed to dynamically learn the feature importance and fine-grained feature interactions. On the one hand, the FiBiNET can dynamically learn the importance of features via the Squeeze-Excitation network (SENET) mechanism; on the other hand, it is able to effectively learn the feature interactions via bilinear function. We conduct extensive experiments on two real-world datasets and show that our shallow model outperforms other shallow models such as factorization machine(FM) and field-aware factorization machine(FFM). In order to improve performance further, we combine a classical deep neural network(DNN) component with the shallow model to be a deep model. The deep FiBiNET consistently outperforms the other state-of-the-art deep models such as DeepFM and extreme deep factorization machine(XdeepFM).

Junlin Zhang, Tongwen Huang, Zhiqi Zhang

Click through rate (CTR) estimation is a fundamental task in personalized advertising and recommender systems. Recent years have witnessed the success of both the deep learning based model and attention mechanism in various tasks in computer vision (CV) and natural language processing (NLP). How to combine the attention mechanism with deep CTR model is a promising direction because it may ensemble the advantages of both sides. Although some CTR model such as Attentional Factorization Machine (AFM) has been proposed to model the weight of second order interaction features, we posit the evaluation of feature importance before explicit feature interaction procedure is also important for CTR prediction tasks because the model can learn to selectively highlight the informative features and suppress less useful ones if the task has many input features. In this paper, we propose a new neural CTR model named Field Attentive Deep Field-aware Factorization Machine (FAT-DeepFFM) by combining the Deep Field-aware Factorization Machine (DeepFFM) with Compose-Excitation network (CENet) field attention mechanism which is proposed by us as an enhanced version of Squeeze-Excitation network (SENet) to highlight the feature importance. We conduct extensive experiments on two real-world datasets and the experiment results show that FAT-DeepFFM achieves the best performance and obtains different improvements over the state-of-the-art methods. We also compare two kinds of attention mechanisms (attention before explicit feature interaction vs. attention after explicit feature interaction) and demonstrate that the former one outperforms the latter one significantly.

Andrei Barbu, Alexander Bridge, Zachary Burchill et al.

We present a system that produces sentential descriptions of video: who did what to whom, and where and how they did it. Action class is rendered as a verb, participant objects as noun phrases, properties of those objects as adjectival modifiers in those noun phrases, spatial relations between those participants as prepositional phrases, and characteristics of the event as prepositional-phrase adjuncts and adverbial modifiers. Extracting the information needed to render these linguistic entities requires an approach to event recognition that recovers object tracks, the trackto-role assignments, and changing body posture.

Andrei Barbu, Alexander Bridge, Dan Coroian et al.

We present an approach to labeling short video clips with English verbs as event descriptions. A key distinguishing aspect of this work is that it labels videos with verbs that describe the spatiotemporal interaction between event participants, humans and objects interacting with each other, abstracting away all object-class information and fine-grained image characteristics, and relying solely on the coarse-grained motion of the event participants. We apply our approach to a large set of 22 distinct verb classes and a corpus of 2,584 videos, yielding two surprising outcomes. First, a classification accuracy of greater than 70% on a 1-out-of-22 labeling task and greater than 85% on a variety of 1-out-of-10 subsets of this labeling task is independent of the choice of which of two different time-series classifiers we employ. Second, we achieve this level of accuracy using a highly impoverished intermediate representation consisting solely of the bounding boxes of one or two event participants as a function of time. This indicates that successful event recognition depends more on the choice of appropriate features that characterize the linguistic invariants of the event classes than on the particular classifier algorithms.

Andrei Barbu, Alexander Bridge, Zachary Burchill et al.

We present a system that produces sentential descriptions of video: who did what to whom, and where and how they did it. Action class is rendered as a verb, participant objects as noun phrases, properties of those objects as adjectival modifiers in those noun phrases,spatial relations between those participants as prepositional phrases, and characteristics of the event as prepositional-phrase adjuncts and adverbial modifiers. Extracting the information needed to render these linguistic entities requires an approach to event recognition that recovers object tracks, the track-to-role assignments, and changing body posture.