Researcher profile

Zhen Zhao

Zhen Zhao contributes to research discovery and scholarly infrastructure.

ResearcherAffiliation not importedOpen to collaborate

Computer Vision Machine Learning Artificial Intelligence astro-ph.IM Computation and Language cond-mat.supr-con eess.IV quant-ph Robotics

Trust snapshot

Quick read

Trust 21 - EmergingVerification L1Unclaimed author

12works

0followers

9topics

4close collaborators

Actions

Decide how to stay connected

Follow researcher0

Claiming links this public author record to a researcher profile and unlocks direct collaboration workflows.

Direct collaboration

Open a focused conversation when the fit is right

Claim this author entity first to unlock direct invitations.

Inspect adjacent work, topics, institutions and collaborators without jumping out to a separate graph page.

Building this graph slice

BZPEER is loading the nearby papers, people, topics and institutions for this page.

preprint2026arXiv

Are Tools Always Beneficial? Learning to Invoke Tools Adaptively for Dual-Mode Multimodal LLM Reasoning

Tool-augmented reasoning has emerged as a promising direction for enhancing the reasoning capabilities of multimodal large language models (MLLMs). However, existing studies mainly focus on enabling models to perform tool invocation, while neglecting the necessity of invoking tools. We argue that tool usage is not always beneficial, as redundant or inappropriate invocations largely increase reasoning overhead and even mislead model predictions. To address this issue, we introduce AutoTool, a model that adaptively decides whether to invoke tools according to the characteristics of each query. Within a reinforcement learning framework, we design an explicit dual-mode reasoning strategy with mode-specific reward functions to guide the model toward producing accurate responses. Moreover, to prevent premature bias toward a single reasoning mode, AutoTool jointly explores and balances tool-assisted and text-centric reasoning throughout training, and promotes free exploration in later stages. Extensive experiments demonstrate that AutoTool exhibits outstanding performance and high efficiency, yielding a 21.8\% accuracy gain on V* benchmark compared to the base model, and a 44.9\% improvement in efficiency over existing tool-augmented methods on POPE benchmark. Code is available at https://github.com/MQinghe/AutoTool.

preprint2026arXiv

CoLVR: Enhancing Exploratory Latent Visual Reasoning via Contrastive Optimization

Due to the potential for exploratory reasoning of Latent Visual Reasoning, recent works tend to enable MLLMs (Multimodal Large Language Models) to perform visual reasoning by propagating continuous hidden states instead of decoding intermediate steps into discrete tokens. However, existing works typically rely on hard alignment objectives to force latent representations to match predefined visual features, thereby severely limiting the exploratory of latent reasoning process. To address this problem, we propose CoLVR (Contrastive Optimization for Latent Visual Reasoning). To obtain a more exploratory visual reasoning, CoLVR introduces a latent contrastive training framework. Firstly, CoLVR learns diverse and exploratory representations with a latent contrastive objective guided by angle-based perturbation, which expands the semantic latent space and avoids over-constrained embedding. Then, CoLVR employs a latent trajectory contrastive reward for RL (Reinforcement Learning) post-training to enable fine-grained optimization of latent visual reasoning process and thus fostering diverse reasoning behaviors. Experiments demonstrate that CoLVR significantly enhances the exploratory capability of latent representations, achieving average improvements of 5.83% on VSP and 8.00% on Jigsaw, while also outperforming existing latent models on out of domain benchmarks, with a 3.40% gain on MMStar. The data, codes, and models are released at https://github.com/Oscar-dzy/CoLVR.

preprint2026arXiv

Semi-MedRef: Semi-Supervised Medical Referring Image Segmentation with Cross-Modal Alignment

Medical referring image segmentation (MRIS) requires pixel-level masks aligned with textual descriptions of anatomical locations, making annotation costly in low-label regimes. Semi-supervised learning (SSL) can mitigate this burden by leveraging unlabeled data, but its success hinges on maintaining reliable image-text alignment under perturbations. Most existing SSL-based referred segmentation methods use either independent or simplistic multi-modal perturbations (e.g., left-right flips), without fully addressing cross-modal alignment under strong augmentation, while CutMix, highly effective in single-modal SSL, remains underexplored in multi-modal settings due to its tendency to disrupt image-text coherence. We propose Semi-MedRef, a teacher-student SSL framework designed to explicitly maintain consistency between medical images and positional language through three alignment-preserving components: T-PatchMix, a cross-modal CutMix-style augmentation that synchronizes patch mixing with referring expressions via position-constrained and probability-driven rules; PosAug, a position-aware text augmentation that masks or fuzzes anatomical phrases; and ITCL, a position-guided image-text contrastive learning module, which leverages positional pseudo-labels to construct soft anatomical positives and strengthen medically grounded cross-modal alignment. Experiments on QaTa-COV19 and MosMedData+ demonstrate that Semi-MedRef consistently outperforms both fully supervised and semi-supervised baselines across all label regimes.

preprint2026arXiv

SOAR: Real-Time Joint Optimization of Order Allocation and Robot Scheduling in Robotic Mobile Fulfillment Systems

Robotic Mobile Fulfillment Systems (RMFS) rely on mobile robots for automated inventory transportation, coordinating order allocation and robot scheduling to enhance warehousing efficiency. However, optimizing RMFS is challenging due to strict real-time constraints and the strong coupling of multi-phase decisions. Existing methods either decompose the problem into isolated sub-tasks to guarantee responsiveness at the cost of global optimality, or rely on computationally expensive global optimization models that are unsuitable for dynamic industrial environments. To bridge this gap, we propose SOAR, a unified Deep Reinforcement Learning framework for real-time joint optimization. SOAR transforms order allocation and robot scheduling into a unified process by utilizing soft order allocations as observations. We formulate this as an Event-Driven Markov Decision Process, enabling the agent to perform simultaneous scheduling in response to asynchronous system events. Technically, we employ a Heterogeneous Graph Transformer to encode the warehouse state and integrate phased domain knowledge. Additionally, we incorporate a reward shaping strategy to address sparse feedback in long-horizon tasks. Extensive experiments on synthetic and real-world industrial datasets, in collaboration with Geekplus, demonstrate that SOAR reduces global makespan by 7.5\% and average order completion time by 15.4\% with sub-100ms latency. Furthermore, sim-to-real deployment confirms its practical viability and significant performance gains in production environments. The code is available at https://github.com/200815147/SOAR.

preprint2022arXiv

Fast Quantum Calibration using Bayesian Optimization with State Parameter Estimator for Non-Markovian Environment

As quantum systems expand in size and complexity, manual qubit characterization and gate optimization will be a non-scalable and time-consuming venture. Physical qubits have to be carefully calibrated because quantum processors are very sensitive to the external environment, with control hardware parameters slowly drifting during operation, affecting gate fidelity. Currently, existing calibration techniques require complex and lengthy measurements to independently control the different parameters of each gate and are unscalable to large quantum systems. Therefore, fully automated protocols with the desired functionalities are required to speed up the calibration process. This paper aims to propose single-qubit calibration of superconducting qubits under continuous weak measurements from a real physical experimental settings point of view. We propose a real-time optimal estimator of qubit states, which utilizes weak measurements and Bayesian optimization to find the optimal control pulses for gate design. Our numerical results demonstrate a significant reduction in the calibration process, obtaining a high gate fidelity. Using the proposed estimator we estimated the qubit state with and without measurement noise and the estimation error between the qubit state and the estimator state is less than 0.02. With this setup, we drive an approximated pi pulse with final fidelity of 0.9928. This shows that our proposed strategy is robust against the presence of measurement and environmental noise and can also be applicable for the calibration of many other quantum computation technologies.

preprint2022arXiv

Titanium-based kagome superconductor CsTi_3Bi_5 and topological states

Since the discovery of a new family of vanadium-based kagome superconductor AV3Sb5 (A=K, Rb, and Cs) with topological band structures, extensive effort has been devoted to exploring the origin of superconducting states and the intertwined orders. Meanwhile, searching for new types of superconductors with novel physical properties and higher superconducting transition temperatures has always been a major thread in the history of superconductor research. Here we report a successful fabrication and the topological states of a Titanium-based kagome metal CsTi3Bi5 (CT3B5) crystal. The as-grown CT3B5 crystal is of high quality and possesses a perfect two-dimensional kagome net of Titanium. The superconductivity of the CT3B5 crystal shows that the critical temperature Tc is of ~4.8 K. First-principle calculations predict that the CT3B5 has robust topological surface states, implying that CT3B5 is a Z2 topological kagome superconductor. This finding provides a new type of superconductors and the base for exploring the origin of superconductivity and topological states in kagome superconductors.

preprint2020arXiv

Active Crowd Counting with Limited Supervision

To learn a reliable people counter from crowd images, head center annotations are normally required. Annotating head centers is however a laborious and tedious process in dense crowds. In this paper, we present an active learning framework which enables accurate crowd counting with limited supervision: given a small labeling budget, instead of randomly selecting images to annotate, we first introduce an active labeling strategy to annotate the most informative images in the dataset and learn the counting model upon them. The process is repeated such that in every cycle we select the samples that are diverse in crowd density and dissimilar to previous selections. In the last cycle when the labeling budget is met, the large amount of unlabeled data are also utilized: a distribution classifier is introduced to align the labeled data with unlabeled data; furthermore, we propose to mix up the distribution labels and latent representations of data in the network to particularly improve the distribution alignment in-between training samples. We follow the popular density estimation pipeline for crowd counting. Extensive experiments are conducted on standard benchmarks i.e. ShanghaiTech, UCF CC 50, MAll, TRANCOS, and DCC. By annotating limited number of images (e.g. 10% of the dataset), our method reaches levels of performance not far from the state of the art which utilize full annotations of the dataset.

preprint2020arXiv

Adaptive Object Detection with Dual Multi-Label Prediction

In this paper, we propose a novel end-to-end unsupervised deep domain adaptation model for adaptive object detection by exploiting multi-label object recognition as a dual auxiliary task. The model exploits multi-label prediction to reveal the object category information in each image and then uses the prediction results to perform conditional adversarial global feature alignment, such that the multi-modal structure of image features can be tackled to bridge the domain divergence at the global feature level while preserving the discriminability of the features. Moreover, we introduce a prediction consistency regularization mechanism to assist object detection, which uses the multi-label prediction results as an auxiliary regularization information to ensure consistent object category discoveries between the object recognition task and the object detection task. Experiments are conducted on a few benchmark datasets and the results show the proposed model outperforms the state-of-the-art comparison methods.

preprint2020arXiv

Ensemble Model with Batch Spectral Regularization and Data Blending for Cross-Domain Few-Shot Learning with Unlabeled Data

In this paper, we present our proposed ensemble model with batch spectral regularization and data blending mechanisms for the Track 2 problem of the cross-domain few-shot learning (CD-FSL) challenge. We build a multi-branch ensemble framework by using diverse feature transformation matrices, while deploying batch spectral feature regularization on each branch to improve the model's transferability. Moreover, we propose a data blending method to exploit the unlabeled data and augment the sparse support set in the target domain. Our proposed model demonstrates effective performance on the CD-FSL benchmark tasks.

preprint2020arXiv

Feature Transformation Ensemble Model with Batch Spectral Regularization for Cross-Domain Few-Shot Classification

In this paper, we propose a feature transformation ensemble model with batch spectral regularization for the Cross-domain few-shot learning (CD-FSL) challenge. Specifically, we proposes to construct an ensemble prediction model by performing diverse feature transformations after a feature extraction network. On each branch prediction network of the model we use a batch spectral regularization term to suppress the singular values of the feature matrix during pre-training to improve the generalization ability of the model. The proposed model can then be fine tuned in the target domain to address few-shot classification. We also further apply label propagation, entropy minimization and data augmentation to mitigate the shortage of labeled data in target domains. Experiments are conducted on a number of CD-FSL benchmark tasks with four target domains and the results demonstrate the superiority of our proposed model.

preprint2020arXiv

Mutual Learning Network for Multi-Source Domain Adaptation

Early Unsupervised Domain Adaptation (UDA) methods have mostly assumed the setting of a single source domain, where all the labeled source data come from the same distribution. However, in practice the labeled data can come from multiple source domains with different distributions. In such scenarios, the single source domain adaptation methods can fail due to the existence of domain shifts across different source domains and multi-source domain adaptation methods need to be designed. In this paper, we propose a novel multi-source domain adaptation method, Mutual Learning Network for Multiple Source Domain Adaptation (ML-MSDA). Under the framework of mutual learning, the proposed method pairs the target domain with each single source domain to train a conditional adversarial domain adaptation network as a branch network, while taking the pair of the combined multi-source domain and target domain to train a conditional adversarial adaptive network as the guidance network. The multiple branch networks are aligned with the guidance network to achieve mutual learning by enforcing JS-divergence regularization over their prediction probability distributions on the corresponding target data. We conduct extensive experiments on multiple multi-source domain adaptation benchmark datasets. The results show the proposed ML-MSDA method outperforms the comparison methods and achieves the state-of-the-art performance.

preprint2019arXiv

A cosmic microscope to probe the Universe from Present to Cosmic Dawn - dual-element low-frequency space VLBI observatory

A space-based very long baseline interferometry (VLBI) programme, named as the Cosmic Microscope, is proposed to involve dual VLBI telescopes in the space working together with giant ground-based telescopes (e.g., Square Kilometre Array, FAST, Arecibo) to image the low radio frequency Universe with the purpose of unraveling the compact structure of cosmic constituents including supermassive black holes and binaries, pulsars, astronomical masers and the underlying source, and exoplanets amongst others. The operational frequency bands are 30, 74, 330 and 1670 MHz, supporting broad science areas. The mission plans to launch two 30-m-diameter radio telescopes into 2,000 km x 90,000 km elliptical orbits. The two telescopes can work in flexibly diverse modes: (i) space-ground VLBI. The maximum space-ground baseline length is about 100,000 km; it provides a high-dynamic-range imaging capacity with unprecedented high resolutions at low frequencies (0.4 mas at 1.67 GHz and 20 mas at 30 MHz) enabling studies of exoplanets and supermassive black hole binaries (which emit nanoHz gravitational waves); (ii) space-space single-baseline VLBI. This unique baseline enables the detection of flaring hydroxyl masers, and more precise position measurement of pulsars and radio transients at milli-arcsecond level; (iii) single dish mode, where each telescope can be used to monitor transient bursts and rapidly trigger follow-up VLBI observations. The large space telescope will also contribute in measuring and constraining the total angular power spectrum from the Epoch of Reionization. In short, the Cosmic Microscope offers astronomers the opportunity to conduct novel, frontier science.

Institution

Affiliation not imported yet

This author record came from a source that does not expose affiliation metadata. Once the author claims the profile or we enrich the record from another provider, this section will link to the concrete institution.

Topic footprint