arXiv subjects

Yifei Sun

Publications and source records attributed to Yifei Sun.

At least 73 records · Page 4Linked to original sources

arXiv2024

Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets

The pre-training of visual representations has enhanced the efficiency of robot learning. Due to the lack of large-scale in-domain robotic datasets, prior works utilize in-the-wild human videos to pre-train robotic visual representation. Despite their promising results, representations from human videos are inevitably subject to distribution shifts and lack the dynamics information crucial for task completion. We first evaluate various pre-trained representations in terms of their correlation to the downstream robotic manipulation tasks (i.e., manipulation centricity). Interestingly, we find that the "manipulation centricity" is a strong indicator of success rates when applied to downstream tasks. Drawing from these findings, we propose Manipulation Centric Representation (MCR), a foundation representation learning framework capturing both visual features and the dynamics information such as actions and proprioceptions of manipulation tasks to improve manipulation centricity. Specifically, we pre-train a visual encoder on the DROID robotic dataset and leverage motion-relevant data such as robot proprioceptive states and actions. We introduce a novel contrastive loss that aligns visual observations with the robot's proprioceptive state-action dynamics, combined with a behavior cloning (BC)-like actor loss to predict actions during pre-training, along with a time contrastive loss. Empirical results across 4 simulation domains with 20 tasks verify that MCR outperforms the strongest baseline method by 14.8%. Moreover, MCR boosts the performance of data-efficient learning with a UR5e arm on 3 real-world tasks by 76.9%. Project website: https://robots-pretrain-robots.github.io/.

cs.RO↗

arXiv2024

EmbSum: Leveraging the Summarization Capabilities of Large Language Models for Content-Based Recommendations

Content-based recommendation systems play a crucial role in delivering personalized content to users in the digital world. In this work, we introduce EmbSum, a novel framework that enables offline pre-computations of users and candidate items while capturing the interactions within the user engagement history. By utilizing the pretrained encoder-decoder model and poly-attention layers, EmbSum derives User Poly-Embedding (UPE) and Content Poly-Embedding (CPE) to calculate relevance scores between users and candidate items. EmbSum actively learns the long user engagement histories by generating user-interest summary with supervision from large language model (LLM). The effectiveness of EmbSum is validated on two datasets from different domains, surpassing state-of-the-art (SoTA) methods with higher accuracy and fewer parameters. Additionally, the model's ability to generate summaries of user interests serves as a valuable by-product, enhancing its usefulness for personalized content recommendations.

cs.IR↗

arXiv2024

Chromosomal Structural Abnormality Diagnosis by Homologous Similarity

Pathogenic chromosome abnormalities are very common among the general population. While numerical chromosome abnormalities can be quickly and precisely detected, structural chromosome abnormalities are far more complex and typically require considerable efforts by human experts for identification. This paper focuses on investigating the modeling of chromosome features and the identification of chromosomes with structural abnormalities. Most existing data-driven methods concentrate on a single chromosome and consider each chromosome independently, overlooking the crucial aspect of homologous chromosomes. In normal cases, homologous chromosomes share identical structures, with the exception that one of them is abnormal. Therefore, we propose an adaptive method to align homologous chromosomes and diagnose structural abnormalities through homologous similarity. Inspired by the process of human expert diagnosis, we incorporate information from multiple pairs of homologous chromosomes simultaneously, aiming to reduce noise disturbance and improve prediction performance. Extensive experiments on real-world datasets validate the effectiveness of our model compared to baselines.

cs.AI↗

arXiv2024

Exploiting Dependency-Aware Priority Adjustment for Mixed-Criticality TSN Flow Scheduling

Time-Sensitive Networking (TSN) serves as a one-size-fits-all solution for mixed-criticality communication, in which flow scheduling is vital to guarantee real-time transmissions. Traditional approaches statically assign priorities to flows based on their associated applications, resulting in significant queuing delays. In this paper, we observe that assigning different priorities to a flow leads to varying delays due to different shaping mechanisms applied to different flow types. Leveraging this insight, we introduce a new scheduling method in mixed-criticality TSN that incorporates a priority adjustment scheme among diverse flow types to mitigate queuing delays and enhance schedulability. Specifically, we propose dependency-aware priority adjustment algorithms tailored to different link-overlapping conditions. Experiments in various settings validate the effectiveness of the proposed method, which enhances the schedulability by 20.57% compared with the SOTA method.

cs.NI↗

arXiv2024

G-Retriever: Retrieval-Augmented Generation for Textual Graph Understanding and Question Answering

Given a graph with textual attributes, we enable users to `chat with their graph': that is, to ask questions about the graph using a conversational interface. In response to a user's questions, our method provides textual replies and highlights the relevant parts of the graph. While existing works integrate large language models (LLMs) and graph neural networks (GNNs) in various ways, they mostly focus on either conventional graph tasks (such as node, edge, and graph classification), or on answering simple graph queries on small or synthetic graphs. In contrast, we develop a flexible question-answering framework targeting real-world textual graphs, applicable to multiple applications including scene graph understanding, common sense reasoning, and knowledge graph reasoning. Toward this goal, we first develop a Graph Question Answering (GraphQA) benchmark with data collected from different tasks. Then, we propose our G-Retriever method, introducing the first retrieval-augmented generation (RAG) approach for general textual graphs, which can be fine-tuned to enhance graph understanding via soft prompting. To resist hallucination and to allow for textual graphs that greatly exceed the LLM's context window size, G-Retriever performs RAG over a graph by formulating this task as a Prize-Collecting Steiner Tree optimization problem. Empirical evaluations show that our method outperforms baselines on textual graph tasks from multiple domains, scales well with larger graph sizes, and mitigates hallucination.~\footnote{Our codes and datasets are available at: \url{https://github.com/XiaoxinHe/G-Retriever}}

cs.LG↗

arXiv2024

SPAR: Personalized Content-Based Recommendation via Long Engagement Attention

Leveraging users' long engagement histories is essential for personalized content recommendations. The success of pretrained language models (PLMs) in NLP has led to their use in encoding user histories and candidate items, framing content recommendations as textual semantic matching tasks. However, existing works still struggle with processing very long user historical text and insufficient user-item interaction. In this paper, we introduce a content-based recommendation framework, SPAR, which effectively tackles the challenges of holistic user interest extraction from the long user engagement history. It achieves so by leveraging PLM, poly-attention layers and attention sparsity mechanisms to encode user's history in a session-based manner. The user and item side features are sufficiently fused for engagement prediction while maintaining standalone representations for both sides, which is efficient for practical model deployment. Moreover, we enhance user profiling by exploiting large language model (LLM) to extract global interests from user engagement history. Extensive experiments on two benchmark datasets demonstrate that our framework outperforms existing state-of-the-art (SoTA) methods.

cs.IR↗

arXiv2024

Exploring Correlations of Self-Supervised Tasks for Graphs

Graph self-supervised learning has sparked a research surge in training informative representations without accessing any labeled data. However, our understanding of graph self-supervised learning remains limited, and the inherent relationships between various self-supervised tasks are still unexplored. Our paper aims to provide a fresh understanding of graph self-supervised learning based on task correlations. Specifically, we evaluate the performance of the representations trained by one specific task on other tasks and define correlation values to quantify task correlations. Through this process, we unveil the task correlations between various self-supervised tasks and can measure their expressive capabilities, which are closely related to downstream performance. By analyzing the correlation values between tasks across various datasets, we reveal the complexity of task correlations and the limitations of existing multi-task learning methods. To obtain more capable representations, we propose Graph Task Correlation Modeling (GraphTCM) to illustrate the task correlations and utilize it to enhance graph self-supervised training. The experimental results indicate that our method significantly outperforms existing methods across various downstream tasks.

cs.LG↗

arXiv2024

Hydrogen-doping mediated solid state thermal switch

Recent reports reveal that isothermal chemical doping of hydrogen in correlated complex oxides such as perovskite nickelates (e.g. NdNiO3) can induce a metal-to-insulator transition (MIT) without the need for temperature modulation. In this work, we interrogate the magnitude change in temperature dependence of thermal conductivity upon chemical doping of hydrogen, as any changes to the thermal properties upon doping offer a route to solid-state thermal switches as well as another potential signal to monitor in a diverse set of sensing, electronic, and optical applications. Using frequency-domain thermoreflectance, we demonstrate that a large concentration of hydrogen (~ 0.1 - 0.5 H/unit cell) completely suppresses the electronic contribution to thermal conductivity in NdNiO3 thin films and reduces the phononic contribution by a factor of 2. These results are critical for the design of next-generation solid-state thermal switches, sensors, extreme environment electronics and neuromorphic memory architectures.

cond-mat.mtrl-sci↗

arXiv2024

Goal-oriented compression for $L_p$-norm-type goal functions: Application to power consumption scheduling

Conventional data compression schemes aim at implementing a trade-off between the rate required to represent the compressed data and the resulting distortion between the original and reconstructed data. However, in more and more applications, what is desired is not reconstruction accuracy but the quality of the realization of a certain task by the receiver. In this paper, the receiver task is modeled by an optimization problem whose parameters have to be compressed by the transmitter. Motivated by applications such as the smart grid, this paper focuses on a goal function which is of $L_p$-norm-type. The aim is to design the precoding, quantization, and decoding stages such that the maximum of the goal function obtained with the compressed version of the parameters is as close as possible to the maximum obtained without compression. The numerical analysis, based on real smart grid signals, clearly shows the benefits of the proposed approach compared to the conventional distortion-based compression paradigm.

eess.SP↗

arXiv2024

Rethinking Grant-Free Protocol in mMTC

This paper revisits the identity detection problem under the current grant-free protocol in massive machine-type communications (mMTC) by asking the following question: for stable identity detection performance, is it enough to permit active devices to transmit preambles without any handshaking with the base station (BS)? Specifically, in the current grant-free protocol, the BS blindly allocates a fixed length of preamble to devices for identity detection as it lacks the prior information on the number of active devices $K$. However, in practice, $K$ varies dynamically over time, resulting in degraded identity detection performance especially when $K$ is large. Consequently, the current grant-free protocol fails to ensure stable identity detection performance. To address this issue, we propose a two-stage communication protocol which consists of estimation of $K$ in Phase I and detection of identities of active devices in Phase II. The preamble length for identity detection in Phase II is dynamically allocated based on the estimated $K$ in Phase I through a table lookup manner such that the identity detection performance could always be better than a predefined threshold. In addition, we design an algorithm for estimating $K$ in Phase I, and exploit the estimated $K$ to reduce the computational complexity of the identity detector in Phase II. Numerical results demonstrate the effectiveness of the proposed two-stage communication protocol and algorithms.

cs.IT↗

arXiv2024

CGS-Mask: Making Time Series Predictions Intuitive for All

Artificial intelligence (AI) has immense potential in time series prediction, but most explainable tools have limited capabilities in providing a systematic understanding of important features over time. These tools typically rely on evaluating a single time point, overlook the time ordering of inputs, and neglect the time-sensitive nature of time series applications. These factors make it difficult for users, particularly those without domain knowledge, to comprehend AI model decisions and obtain meaningful explanations. We propose CGS-Mask, a post-hoc and model-agnostic cellular genetic strip mask-based saliency approach to address these challenges. CGS-Mask uses consecutive time steps as a cohesive entity to evaluate the impact of features on the final prediction, providing binary and sustained feature importance scores over time. Our algorithm optimizes the mask population iteratively to obtain the optimal mask in a reasonable time. We evaluated CGS-Mask on synthetic and real-world datasets, and it outperformed state-of-the-art methods in elucidating the importance of features over time. According to our pilot user study via a questionnaire survey, CGS-Mask is the most effective approach in presenting easily understandable time series prediction results, enabling users to comprehend the decision-making process of AI models with ease.

cs.AI↗

arXiv2024

System and Method to Determine ME/CFS and Long COVID Disease Severity Using a Wearable Sensor

Objective: We present a simple parameter, calculated from a single wearable sensor, that can be used to objectively measure disease severity in people with myalgic encephalomyelitis/chronic fatigue syndrome (ME/CFS) or Long COVID. We call this parameter UpTime. Methods: Prior research has shown that the amount of time a person spends upright, defined as lower legs vertical with feet on the floor, correlates strongly with ME/CFS disease severity. We use a single commercial inertial measurement unit (IMU) attached to the ankle to calculate the percentage of time each day that a person spends upright (i.e., UpTime) and number of Steps/Day. As Long COVID shares symptoms with ME/CFS, we also apply this method to determine Long COVID disease severity. We performed a trial with 55 subjects broken into three cohorts, healthy controls, ME/CFS, and Long COVID. Subjects wore the IMU on their ankle for a period of 7 days. UpTime and Steps/Day were calculated each day and results compared between cohorts. Results: UpTime effectively distinguishes between healthy controls and subjects diagnosed with ME/CFS ($\mathbf{p = 0.00004}$) and between healthy controls and subjects diagnosed with Long COVID ($\mathbf{p = 0.01185}$). Steps/Day did distinguish between controls and subjects with ME/CFS ($\mathbf{p = 0.01}$) but did not distinguish between controls and subjects with Long COVID ($\mathbf{p = 0.3}$). Conclusion: UpTime is an objective measure of ME/CFS and Long COVID severity. UpTime can be used as an objective outcome measure in clinical research and treatment trials. Significance: Objective assessment of ME/CFS and Long COVID disease severity using UpTime could spur development of treatments by enabling the effect of those treatments to be easily measured.

q-bio.QM↗

arXiv2024

Unraveling the nature of quasi van der Waals Epitaxy of magnetic topological insulators Cr: (BixSb1-x)2Te3 on a GaAs (111) substrate through coherently strained interface

Quasi van der Waals Epitaxy (qvdWE) has been realized for decades at the interfaces between 3D and 2D materials or van der Waals materials. The growth of magnetic topological insulators (MTI) Cr: (BixSb1-x)2Te3 (CBST) on GaAs (111) substrates for Quantum Anomalous Hall Effect (QAH) is actually one of the examples of qvdWE, which is not well noticed despite the fact that its advantages have been used in growth of various MTI materials. This is distinguished from the growth of MTIs on other substrates. Although the qvdWE mode has been used in many 2D growth on III-V substrates, the specific features and mechanisms are not well demonstrated and summarized yet. Here in this work, we have for the first time shown the features of both coherent interfaces and the existence of strain originating from qvdWE at the same time.

cond-mat.mtrl-sci↗

arXiv2024

Robust deep labeling of radiological emphysema subtypes using squeeze and excitation convolutional neural networks: The MESA Lung and SPIROMICS Studies

Pulmonary emphysema, the progressive, irreversible loss of lung tissue, is conventionally categorized into three subtypes identifiable on pathology and on lung computed tomography (CT) images. Recent work has led to the unsupervised learning of ten spatially-informed lung texture patterns (sLTPs) on lung CT, representing distinct patterns of emphysematous lung parenchyma based on both textural appearance and spatial location within the lung, and which aggregate into 6 robust and reproducible CT Emphysema Subtypes (CTES). Existing methods for sLTP segmentation, however, are slow and highly sensitive to changes in CT acquisition protocol. In this work, we present a robust 3-D squeeze-and-excitation CNN for supervised classification of sLTPs and CTES on lung CT. Our results demonstrate that this model achieves accurate and reproducible sLTP segmentation on lung CTscans, across two independent cohorts and independently of scanner manufacturer and model.

cs.CV↗

arXiv2024

BS-Diff: Effective Bone Suppression Using Conditional Diffusion Models from Chest X-Ray Images

Chest X-rays (CXRs) are commonly utilized as a low-dose modality for lung screening. Nonetheless, the efficacy of CXRs is somewhat impeded, given that approximately 75% of the lung area overlaps with bone, which in turn hampers the detection and diagnosis of diseases. As a remedial measure, bone suppression techniques have been introduced. The current dual-energy subtraction imaging technique in the clinic requires costly equipment and subjects being exposed to high radiation. To circumvent these issues, deep learning-based image generation algorithms have been proposed. However, existing methods fall short in terms of producing high-quality images and capturing texture details, particularly with pulmonary vessels. To address these issues, this paper proposes a new bone suppression framework, termed BS-Diff, that comprises a conditional diffusion model equipped with a U-Net architecture and a simple enhancement module to incorporate an autoencoder. Our proposed network cannot only generate soft tissue images with a high bone suppression rate but also possesses the capability to capture fine image details. Additionally, we compiled the largest dataset since 2010, including data from 120 patients with high-definition, high-resolution paired CXRs and soft tissue images collected by our affiliated hospital. Extensive experiments, comparative analyses, ablation studies, and clinical evaluations indicate that the proposed BS-Diff outperforms several bone-suppression models across multiple metrics. Our code can be accessed at https://github.com/Benny0323/BS-Diff.

eess.IV↗

arXiv2024

Unsupervised Airway Tree Clustering with Deep Learning: The Multi-Ethnic Study of Atherosclerosis (MESA) Lung Study

High-resolution full lung CT scans now enable the detailed segmentation of airway trees up to the 6th branching generation. The airway binary masks display very complex tree structures that may encode biological information relevant to disease risk and yet remain challenging to exploit via traditional methods such as meshing or skeletonization. Recent clinical studies suggest that some variations in shape patterns and caliber of the human airway tree are highly associated with adverse health outcomes, including all-cause mortality and incident COPD. However, quantitative characterization of variations observed on CT segmented airway tree remain incomplete, as does our understanding of the clinical and developmental implications of such. In this work, we present an unsupervised deep-learning pipeline for feature extraction and clustering of human airway trees, learned directly from projections of 3D airway segmentations. We identify four reproducible and clinically distinct airway sub-types in the MESA Lung CT cohort.

eess.IV↗

arXiv2024

Accurate Leukocyte Detection Based on Deformable-DETR and Multi-Level Feature Fusion for Aiding Diagnosis of Blood Diseases

In standard hospital blood tests, the traditional process requires doctors to manually isolate leukocytes from microscopic images of patients' blood using microscopes. These isolated leukocytes are then categorized via automatic leukocyte classifiers to determine the proportion and volume of different types of leukocytes present in the blood samples, aiding disease diagnosis. This methodology is not only time-consuming and labor-intensive, but it also has a high propensity for errors due to factors such as image quality and environmental conditions, which could potentially lead to incorrect subsequent classifications and misdiagnosis. To address these issues, this paper proposes an innovative method of leukocyte detection: the Multi-level Feature Fusion and Deformable Self-attention DETR (MFDS-DETR). To tackle the issue of leukocyte scale disparity, we designed the High-level Screening-feature Fusion Pyramid (HS-FPN), enabling multi-level fusion. This model uses high-level features as weights to filter low-level feature information via a channel attention module and then merges the screened information with the high-level features, thus enhancing the model's feature expression capability. Further, we address the issue of leukocyte feature scarcity by incorporating a multi-scale deformable self-attention module in the encoder and using the self-attention and cross-deformable attention mechanisms in the decoder, which aids in the extraction of the global features of the leukocyte feature maps. The effectiveness, superiority, and generalizability of the proposed MFDS-DETR method are confirmed through comparisons with other cutting-edge leukocyte detection models using the private WBCDD, public LISC and BCCD datasets. Our source code and private WBCCD dataset are available at https://github.com/JustlfC03/MFDS-DETR.

cs.CV↗

arXiv2023

Generalizable Visual Reinforcement Learning with Segment Anything Model

Learning policies that can generalize to unseen environments is a fundamental challenge in visual reinforcement learning (RL). While most current methods focus on acquiring robust visual representations through auxiliary supervision, pre-training, or data augmentation, the potential of modern vision foundation models remains underleveraged. In this work, we introduce Segment Anything Model for Generalizable visual RL (SAM-G), a novel framework that leverages the promptable segmentation ability of Segment Anything Model (SAM) to enhance the generalization capabilities of visual RL agents. We utilize image features from DINOv2 and SAM to find correspondence as point prompts to SAM, and then SAM produces high-quality masked images for agents directly. Evaluated across 8 DMControl tasks and 3 Adroit tasks, SAM-G significantly improves the visual generalization ability without altering the RL agents' architecture but merely their observations. Notably, SAM-G achieves 44% and 29% relative improvements on the challenging video hard setting on DMControl and Adroit respectively, compared to state-of-the-art methods. Video and code: https://yanjieze.com/SAM-G/

cs.LG↗