Qiang Guan

Cheng Han, Qifan Wang, Sohail A. Dianat et al.

Transformer-based architectures have become the de-facto standard models for diverse vision tasks owing to their superior performance. As the size of the models continues to scale up, model distillation becomes extremely important in various real applications, particularly on devices limited by computational resources. However, prevailing knowledge distillation methods exhibit diminished efficacy when confronted with a large capacity gap between the teacher and the student, e.g, 10x compression rate. In this paper, we present a novel approach named Automatic Multi-step Distillation (AMD) for large-scale vision model compression. In particular, our distillation process unfolds across multiple steps. Initially, the teacher undergoes distillation to form an intermediate teacher-assistant model, which is subsequently distilled further to the student. An efficient and effective optimization framework is introduced to automatically identify the optimal teacher-assistant that leads to the maximal student performance. We conduct extensive experiments on multiple image classification datasets, including CIFAR-10, CIFAR-100, and ImageNet. The findings consistently reveal that our approach outperforms several established baselines, paving a path for future knowledge distillation methods on large-scale vision models.

Liqi Yan, Qifan Wang, Junhan Zhao et al.

First-Person-View (FPV) holds immense potential for revolutionizing the trajectory of Unmanned Aerial Vehicles (UAVs), offering an exhilarating avenue for navigating complex building structures. Yet, traditional Neural Radiance Field (NeRF) methods face challenges such as sampling single points per iteration and requiring an extensive array of views for supervision. UAV videos exacerbate these issues with limited viewpoints and significant spatial scale variations, resulting in inadequate detail rendering across diverse scales. In response, we introduce FPV-NeRF, addressing these challenges through three key facets: (1) Temporal consistency. Leveraging spatio-temporal continuity ensures seamless coherence between frames; (2) Global structure. Incorporating various global features during point sampling preserves space integrity; (3) Local granularity. Employing a comprehensive framework and multi-resolution supervision for multi-scale scene feature representation tackles the intricacies of UAV video spatial scales. Additionally, due to the scarcity of publicly available FPV videos, we introduce an innovative view synthesis method using NeRF to generate FPV perspectives from UAV footage, enhancing spatial perception for drones. Our novel dataset spans diverse trajectories, from outdoor to indoor environments, in the UAV domain, differing significantly from traditional NeRF scenarios. Through extensive experiments encompassing both interior and exterior building structures, FPV-NeRF demonstrates a superior understanding of the UAV flying space, outperforming state-of-the-art methods in our curated UAV dataset. Explore our project page for further insights: https://fpv-nerf.github.io/.

Runjia Zeng, Qifan Wang, Qiang Guan et al.

Fine tuning has been regarded as a de facto approach for adapting large language models (LLMs) to downstream tasks, but the high training memory consumption inherited from LLMs makes this process inefficient. Among existing memory efficient approaches, activation-related optimization has proven particularly effective, as activations consistently dominate overall memory consumption. Although prior arts offer various activation optimization strategies, their data-agnostic nature ultimately results in ineffective and unstable fine tuning. In this paper, we propose TokenSeek, a universal plugin solution for various transformer-based models through instance-aware token seeking and ditching, achieving significant fine-tuning memory savings (e.g., requiring only 14.8% of the memory on Llama3.2 1B) with on-par or even better performance. Furthermore, our interpretable token seeking process reveals the underlying reasons for its effectiveness, offering valuable insights for future research on token efficiency. Homepage: https://runjia.tech/iclr_tokenseek/

Pegah Ahadian, Mingrui Yang, Sixu Chen et al.

Knee osteoarthritis frequently exhibits discordance between structural damage observed in imaging and patient-reported symptoms such as pain. This mismatch complicates clinical interpretation and patient stratification and remains insufficiently modeled in existing decision support systems. We propose a discordance aware multimodal framework that combines machine learning prediction models with a tool grounded multi agent reasoning system. Using baseline data from the FNIH Osteoarthritis Biomarkers Consortium, we trained multimodal models to predict two progression tasks, joint space loss only progression versus non progression, and pain only progression versus non progression. The predictive system integrates three modality specific experts: a CatBoost tabular model using demographic, radiographic, MRI-derived scalar, and biomarker features; MRI image embeddings extracted using a ResNet18 backbone; and Xray embeddings derived from the same architecture. Expert predictions are fused using a stacking ensemble. Residual based models estimate expected pain from structural features, enabling the computation of a pain structure discordance score between observed and expected symptoms. A multi-agent reasoning layer interprets these signals to assign clinically interpretable OA phenotypes and generate phenotype specific management recommendations.

Xinyan Cai, Shiguang Wu, Dafeng Chi et al.

In complex embodied long-horizon manipulation tasks, effective task decomposition and execution require synergistic integration of textual logical reasoning and visual-spatial imagination to ensure efficient and accurate operation. Current methods fail to adopt a unified generation framework for multimodal planning, lead to inconsistent in multimodal planning. To address this challenge, we present \textbf{EVLP (Embodied Vision-Language Planner)}, an innovative multimodal unified generation framework that jointly models linguistic reasoning and visual generation. Our approach achieves multimodal planning for long-horizon tasks through a novel training pipeline incorporating dynamic pretraining and reinforced alignment. Our core innovations consist of three key components: \textbf{1) Unified Multimodal Generation Framework}: For understanding, We integrate semantic information with spatial features to provide comprehensive visual perception. For generation, we directly learn the joint distribution of discrete images for one-step visual synthesis, enabling coordinated language-visual modeling through learnable cross-modal attention mechanisms. \textbf{2) Dynamic Perception Pretraining}: We propose a bidirectional dynamic alignment strategy employing inverse dynamics tasks and forward dynamics tasks, effectively strengthening multimodal correlations within a unified feature space. \textbf{3) Reinforced Supervised Fine-Tuning}: While conducting instruction-based fine-tuning in the unified generation space, we construct a reinforce loss to align the spatial logic between textual actions and generated images, enabling the model to acquire spatio-awared multimodal planning capabilities.

Keer Yang, Guanqun Zhang, Chuan Bi et al.

In recent years, there have been quite a few attempts to apply intelligent techniques to financial trading, i.e., constructing automatic and intelligent trading framework based on historical stock price. Due to the unpredictable, uncertainty and volatile nature of financial market, researchers have also resorted to deep learning to construct the intelligent trading framework. In this paper, we propose to use CNN as the core functionality of such framework, because it is able to learn the spatial dependency (i.e., between rows and columns) of the input data. However, different with existing deep learning-based trading frameworks, we develop novel normalization process to prepare the stock data. In particular, we first empirically observe that the stock data is intrinsically heterogeneous and bursty, and then validate the heterogeneity and burst nature of stock data from a statistical perspective. Next, we design the data normalization method in a way such that the data heterogeneity is preserved and bursty events are suppressed. We verify out developed CNN-based trading framework plus our new normalization method on 29 stocks. Experiment results show that our approach can outperform other comparing approaches.

Shaolun Ruan, Feng Liang, Rohan Ramakrishna et al.

Quantum Neural Networks (QNNs) represent a promising fusion of quantum computing and neural network architectures, offering speed-ups and efficient processing of high-dimensional, entangled data. A crucial component of QNNs is the encoder, which maps classical input data into quantum states. However, choosing suitable encoders remains a significant challenge, largely due to the lack of systematic guidance and the trial-and-error nature of current approaches. This process is further impeded by two key challenges: (1) the difficulty in evaluating encoded quantum states prior to training, and (2) the lack of intuitive methods for analyzing an encoder's ability to effectively distinguish data features. To address these issues, we introduce a novel visualization tool, XQAI-Eyes, which enables QNN developers to compare classical data features with their corresponding encoded quantum states and to examine the mixed quantum states across different classes. By bridging classical and quantum perspectives, XQAI-Eyes facilitates a deeper understanding of how encoders influence QNN performance. Evaluations across diverse datasets and encoder designs demonstrate XQAI-Eyes's potential to support the exploration of the relationship between encoder design and QNN effectiveness, offering a holistic and transparent approach to optimizing quantum encoders. Moreover, domain experts used XQAI-Eyes to derive two key practices for quantum encoder selection, grounded in the principles of pattern preservation and feature mapping.

Yanchuan Tang, Taowen Wang, Yuefei Chen et al.

Vision-Language-Action (VLA) models enable general-purpose robotic policies by mapping visual observations and language instructions to low-level actions, but they often lack reliable introspection. A common practice is to compute a token-level uncertainty signal and take its mean over a rollout. However, mean aggregation can dilute short-lived but safety-critical uncertainty spikes in continuous control. In particular, successful rollouts may contain localized high-entropy segments due to benign noise or non-critical micro-adjustments, while failure rollouts can appear low-entropy for most timesteps and only exhibit brief spikes near the onset of failure. We propose a unified uncertainty quantification approach for predicting rollout success versus failure that (1) uses max-based sliding window pooling to preserve transient risk signals, (2) applies motion-aware stability weighting to emphasize high-frequency action oscillations associated with unstable behaviors, and (3) performs DoF-adaptive calibration via Bayesian Optimization to prioritize kinematically critical axes. Experiments on the LIBERO benchmark show that our method substantially improves failure prediction accuracy and yields more reliable signals for failure detection, which can support downstream human-in-the-loop interventions.

Runjia Zeng, Priyabrata Senapati, Ruixiang Tang et al.

Large language models have recently shown potential in bridging the gap between classical machine learning and quantum machine learning. However, the lack of standardized, high-quality datasets and robust translation frameworks limits progress in this domain. We introduce Q-Bridge, an LLM-guided code translation framework that systematically converts CML implementations into executable QML variants. Our approach builds on a self-involving pipeline that iteratively expands a verified seed codebase into a large-scale dataset, CML-2-QML, integrating verifiable and unverifiable code pairs. The Q-Bridge model is fine-tuned using supervised LoRA adaptation for scalable and memory-efficient training, achieving faithful and interpretable quantum code generation across diverse architectures. Empirical analysis confirms the feasibility of direct CML-to-QML translation and reveals consistent structural alignment between classical and quantum paradigms. Case studies further demonstrate that Q-Bridge can maintain deterministic correctness and also enable creative architectural exploration. This work establishes the first reproducible framework and dataset for LLM-driven quantum code translation, offering a foundation for scalable quantum AI development.

Pegah Ahadian, Wei Xu, Sherry Wang et al.

Sleep disorders have a major impact on both lifestyle and health. Effective sleep disorder prediction from lifestyle and physiological data can provide essential details for early intervention. This research utilizes three deep time series models and facilitates them with explainability approaches for sleep disorder prediction. Specifically, our approach adopts Temporal Convolutional Networks (TCN), Long Short-Term Memory (LSTM) for time series data analysis, and Temporal Fusion Transformer model (TFT). Meanwhile, the temporal attention mechanism and counterfactual explanation with SHapley Additive exPlanations (SHAP) approach are employed to ensure dependable, accurate, and interpretable predictions. Finally, using a large dataset of sleep health measures, our evaluation demonstrates the effect of our method in predicting sleep disorders.

Pegah Ahadian, Yunhe Feng, Karl Kosko et al.

Mathematics education, a crucial and basic field, significantly influences students' learning in related subjects and their future careers. Utilizing artificial intelligence to interpret and comprehend math problems in education is not yet fully explored. This is due to the scarcity of quality datasets and the intricacies of processing handwritten information. In this paper, we present a novel contribution to the field of mathematics education through the development of MNIST-Fraction, a dataset inspired by the renowned MNIST, specifically tailored for the recognition and understanding of handwritten math fractions. Our approach is the utilization of deep learning, specifically Convolutional Neural Networks (CNNs), for the recognition and understanding of handwritten math fractions to effectively detect and analyze fractions, along with their numerators and denominators. This capability is pivotal in calculating the value of fractions, a fundamental aspect of math learning. The MNIST-Fraction dataset is designed to closely mimic real-world scenarios, providing a reliable and relevant resource for AI-driven educational tools. Furthermore, we conduct a comprehensive comparison of our dataset with the original MNIST dataset using various classifiers, demonstrating the effectiveness and versatility of MNIST-Fraction in both detection and classification tasks. This comparative analysis not only validates the practical utility of our dataset but also offers insights into its potential applications in math education. To foster collaboration and further research within the computational and educational communities. Our work aims to bridge the gap in high-quality educational resources for math learning, offering a valuable tool for both educators and researchers in the field.

Hailong Jiang, Jianfeng Zhu, Yao Wan et al.

Intermediate Representations (IRs) play a critical role in compiler design and program analysis, yet their comprehension by Large Language Models (LLMs) remains underexplored. In this paper, we present an explorative empirical study evaluating the capabilities of six state-of-the-art LLMs: GPT-4, GPT-3, DeepSeek, Gemma 2, Llama 3, and Code Llama, in understanding IRs. Specifically, we assess model performance across four core tasks: control flow graph reconstruction, decompilation, code summarization, and execution reasoning. While LLMs exhibit competence in parsing IR syntax and identifying high-level structures, they consistently struggle with instruction-level reasoning, especially in control flow reasoning, loop handling, and dynamic execution. Common failure modes include misinterpreting branching instructions, omitting critical operations, and relying on heuristic reasoning rather than precise instruction-level logic. Our findings highlight the need for IR-specific enhancements in LLM design. We recommend fine-tuning on structured IR datasets and integrating control-flow-sensitive architectures to improve model effectiveness. All experimental data and source code are publicly available at

Guilin Zhang, Wulan Guo, Ziqi Tan et al.

Quantum machine learning (QML) promises compact and expressive representations, but suffers from the measurement bottleneck - a narrow quantum-to-classical readout that limits performance and amplifies privacy risk. We propose a lightweight residual hybrid architecture that concatenates quantum features with raw inputs before classification, bypassing the bottleneck without increasing quantum complexity. Experiments show our model outperforms pure quantum and prior hybrid models in both centralized and federated settings. It achieves up to +55% accuracy improvement over quantum baselines, while retaining low communication cost and enhanced privacy robustness. Ablation studies confirm the effectiveness of the residual connection at the quantum-classical interface. Our method offers a practical, near-term pathway for integrating quantum models into privacy-sensitive, resource-constrained settings like federated edge learning.

Cheng Han, Yawen Lu, Guohao Sun et al.

In this work, we introduce the Prototypical Transformer (ProtoFormer), a general and unified framework that approaches various motion tasks from a prototype perspective. ProtoFormer seamlessly integrates prototype learning with Transformer by thoughtfully considering motion dynamics, introducing two innovative designs. First, Cross-Attention Prototyping discovers prototypes based on signature motion patterns, providing transparency in understanding motion scenes. Second, Latent Synchronization guides feature representation learning via prototypes, effectively mitigating the problem of motion uncertainty. Empirical results demonstrate that our approach achieves competitive performance on popular motion tasks such as optical flow and scene depth. Furthermore, it exhibits generality across various downstream tasks, including object tracking and video stabilization.

Shaolun Ruan, Yong Wang, Qiang Guan

Soft errors have become one of the major concerns for HPC applications, as those errors can result in seriously corrupted outcomes, such as silent data corruptions (SDCs). Prior studies on error resilience have studied the robustness of HPC applications. However, it is still difficult for program developers to identify potential vulnerability to soft errors. In this paper, we present Visilence, a novel visualization tool to visually analyze error vulnerability based on the control-flow graph generated from HPC applications. Visilence efficiently visualizes the affected program states under injected errors and presents the visual analysis of the most vulnerable parts of an application. We demonstrate the effectiveness of Visilence through a case study.

Shaolun Ruan, Yong Wang, Hailong Jiang et al.

Cloud systems are becoming increasingly powerful and complex. It is highly challenging to identify anomalous execution behaviors and pinpoint problems by examining the overwhelming intermediate results/states in complex application workflows. Domain scientists urgently need a friendly and functional interface to understand the quality of the computing services and the performance of their applications in real time. To meet these needs, we explore data generated by job schedulers and investigate general performance metrics (e.g., utilization of CPU, memory and disk I/O). Specifically, we propose an interactive visual analytics approach, BatchLens, to provide both providers and users of cloud service with an intuitive and effective way to explore the status of system batch jobs and help them conduct root-cause analysis of anomalous behaviors in batch jobs. We demonstrate the effectiveness of BatchLens through a case study on the public Alibaba bench workload trace datasets.

Shaolun Ruan, Yong Wang, Qiang Guan

State change comparison of multiple data items is often necessary in multiple application domains, such as medical science, financial engineering, sociology, biological science, etc. Slope graphs and grouped bar charts have been widely used to show a "before-and-after" story of different data states and indicate their changes. However, they visualize state changes as either slope or difference of bars, which has been proved less effective for quantitative comparison. Also, both visual designs suffer from visual clutter issues with an increasing number of data items. In this paper, we propose Intercept Graph, a novel visual design to facilitate effective interactive comparison of state changes. Specifically, a radial design is proposed to visualize the starting and ending states of each data item and the line segment length explicitly encodes the "state change". By interactively adjusting the radius of the inner circular axis, Intercept Graph can smoothly filter the large state changes and magnify the difference between similar state changes, mitigating the visual clutter issues and enhancing the effective comparison of state changes. We conducted a case study through comparing Intercept Graph with slope graphs and grouped bar charts on real datasets to demonstrate the effectiveness of Intercept Graph.

Yuxin Yang, Xitong Zhang, Qiang Guan et al.

Deep learning and data-driven approaches have shown great potential in scientific domains. The promise of data-driven techniques relies on the availability of a large volume of high-quality training datasets. Due to the high cost of obtaining data through expensive physical experiments, instruments, and simulations, data augmentation techniques for scientific applications have emerged as a new direction for obtaining scientific data recently. However, existing data augmentation techniques originating from computer vision, yield physically unacceptable data samples that are not helpful for the domain problems that we are interested in. In this paper, we develop new data augmentation techniques based on convolutional neural networks. Specifically, our generative models leverage different physics knowledge (such as governing equations, observable perception, and physics phenomena) to improve the quality of the synthetic data. To validate the effectiveness of our data augmentation techniques, we apply them to solve a subsurface seismic full-waveform inversion using simulated CO$_2$ leakage data. Our interest is to invert for subsurface velocity models associated with very small CO$_2$ leakage. We validate the performance of our methods using comprehensive numerical tests. Via comparison and analysis, we show that data-driven seismic imaging can be significantly enhanced by using our data augmentation techniques. Particularly, the imaging quality has been improved by 15% in test scenarios of general-sized leakage and 17% in small-sized leakage when using an augmented training set obtained with our techniques.

Xinyu Chen, Qiang Guan, Li-Ta Lo et al.

Convolutional Neural Networks(CNNs) are complex systems. They are trained so they can adapt their internal connections to recognize images, texts and more. It is both interesting and helpful to visualize the dynamics within such deep artificial neural networks so that people can understand how these artificial networks are learning and making predictions. In the field of scientific simulations, visualization tools like Paraview have long been utilized to provide insights and understandings. We present in situ TensorView to visualize the training and functioning of CNNs as if they are systems of scientific simulations. In situ TensorView is a loosely coupled in situ visualization open framework that provides multiple viewers to help users to visualize and understand their networks. It leverages the capability of co-processing from Paraview to provide real-time visualization during training and predicting phases. This avoid heavy I/O overhead for visualizing large dynamic systems. Only a small number of lines of codes are injected in TensorFlow framework. The visualization can provide guidance to adjust the architecture of networks, or compress the pre-trained networks. We showcase visualizing the training of LeNet-5 and VGG16 using in situ TensorView.