Search arXiv⌕ Search

arXiv subjects

Yu Gu

Publications and source records attributed to Yu Gu.

At least 181 records · Page 10Linked to original sources
arXiv2023

BiomedJourney: Counterfactual Biomedical Image Generation by Instruction-Learning from Multimodal Patient Journeys

Rapid progress has been made in instruction-learning for image editing with natural-language instruction, as exemplified by InstructPix2Pix. In biomedicine, such methods can be applied to counterfactual image generation, which helps differentiate causal structure from spurious correlation and facilitate robust image interpretation for disease progression modeling. However, generic image-editing models are ill-suited for the biomedical domain, and counterfactual biomedical image generation is largely underexplored. In this paper, we present BiomedJourney, a novel method for counterfactual biomedical image generation by instruction-learning from multimodal patient journeys. Given a patient with two biomedical images taken at different time points, we use GPT-4 to process the corresponding imaging reports and generate a natural language description of disease progression. The resulting triples (prior image, progression description, new image) are then used to train a latent diffusion model for counterfactual biomedical image generation. Given the relative scarcity of image time series data, we introduce a two-stage curriculum that first pretrains the denoising network using the much more abundant single image-report pairs (with dummy prior image), and then continues training using the counterfactual triples. Experiments using the standard MIMIC-CXR dataset demonstrate the promise of our method. In a comprehensive battery of tests on counterfactual medical image generation, BiomedJourney substantially outperforms prior state-of-the-art methods in instruction image editing and medical image generation such as InstructPix2Pix and RoentGen. To facilitate future study in counterfactual medical generation, we plan to release our instruction-learning code and pretrained models.

cs.CV↗
arXiv2023

ONNXExplainer: an ONNX Based Generic Framework to Explain Neural Networks Using Shapley Values

Understanding why a neural network model makes certain decisions can be as important as the inference performance. Various methods have been proposed to help practitioners explain the prediction of a neural network model, of which Shapley values are most popular. SHAP package is a leading implementation of Shapley values to explain neural networks implemented in TensorFlow or PyTorch but lacks cross-platform support, one-shot deployment and is highly inefficient. To address these problems, we present the ONNXExplainer, which is a generic framework to explain neural networks using Shapley values in the ONNX ecosystem. In ONNXExplainer, we develop its own automatic differentiation and optimization approach, which not only enables One-Shot Deployment of neural networks inference and explanations, but also significantly improves the efficiency to compute explanation with less memory consumption. For fair comparison purposes, we also implement the same optimization in TensorFlow and PyTorch and measure its performance against the current state of the art open-source counterpart, SHAP. Extensive benchmarks demonstrate that the proposed optimization approach improves the explanation latency of VGG19, ResNet50, DenseNet201, and EfficientNetB0 by as much as 500%.

cs.LG↗
arXiv2023

DurIAN-E: Duration Informed Attention Network For Expressive Text-to-Speech Synthesis

This paper introduces an improved duration informed attention neural network (DurIAN-E) for expressive and high-fidelity text-to-speech (TTS) synthesis. Inherited from the original DurIAN model, an auto-regressive model structure in which the alignments between the input linguistic information and the output acoustic features are inferred from a duration model is adopted. Meanwhile the proposed DurIAN-E utilizes multiple stacked SwishRNN-based Transformer blocks as linguistic encoders. Style-Adaptive Instance Normalization (SAIN) layers are exploited into frame-level encoders to improve the modeling ability of expressiveness. A denoiser incorporating both denoising diffusion probabilistic model (DDPM) for mel-spectrograms and SAIN modules is conducted to further improve the synthetic speech quality and expressiveness. Experimental results prove that the proposed expressive TTS model in this paper can achieve better performance than the state-of-the-art approaches in both subjective mean opinion score (MOS) and preference tests.

eess.AS↗
arXiv2023

SOT-MRAM-Enabled Probabilistic Binary Neural Networks for Noise-Tolerant and Fast Training

We report the use of spin-orbit torque (SOT) magnetoresistive random-access memory (MRAM) to implement a probabilistic binary neural network (PBNN) for resource-saving applications. The in-plane magnetized SOT (i-SOT) MRAM not only enables field-free magnetization switching with high endurance (> 10^11), but also hosts multiple stable probabilistic states with a low device-to-device variation (< 6.35%). Accordingly, the proposed PBNN outperforms other neural networks by achieving an 18* increase in training speed, while maintaining an accuracy above 97% under the write and read noise perturbations. Furthermore, by applying the binarization process with an additional SOT-MRAM dummy module, we demonstrate an on-chip MNIST inference performance close to the ideal baseline using our SOT-PBNN hardware.

physics.app-ph↗
arXiv2023

Digital Twin System for Home Service Robot Based on Motion Simulation

In order to improve the task execution capability of home service robot, and to cope with the problem that purely physical robot platforms cannot sense the environment and make decisions online, a method for building digital twin system for home service robot based on motion simulation is proposed. A reliable mapping of the home service robot and its working environment from physical space to digital space is achieved in three dimensions: geometric, physical and functional. In this system, a digital space-oriented URDF file parser is designed and implemented for the automatic construction of the robot geometric model. Next, the physical model is constructed from the kinematic equations of the robot and an improved particle swarm optimization algorithm is proposed for the inverse kinematic solution. In addition, to adapt to the home environment, functional attributes are used to describe household objects, thus improving the semantic description of the digital space for the real home environment. Finally, through geometric model consistency verification, physical model validity verification and virtual-reality consistency verification, it shows that the digital twin system designed in this paper can construct the robot geometric model accurately and completely, complete the operation of household objects successfully, and the digital twin system is effective and practical.

cs.RO↗
arXiv2023

Observation of integer and fractional quantum anomalous Hall effects in twisted bilayer MoTe2

The interplay between strong correlations and topology can lead to the emergence of intriguing quantum states of matter. One well-known example is the fractional quantum Hall effect, where exotic electron fluids with fractionally charged excitations form in partially filled Landau levels. The emergence of topological moiré flat bands provides exciting opportunities to realize the lattice analogs of both the integer and fractional quantum Hall states without the need for an external magnetic field. These states are known as the integer and fractional quantum anomalous Hall (IQAH and FQAH) states. Here, we present direct transport evidence of the existence of both IQAH and FQAH states in twisted bilayer MoTe2 (AA stacked). At zero magnetic field, we observe well-quantized Hall resistance of h/e2 around moiré filling factor ν = -1 (corresponding to one hole per moiré unit cell), and nearly-quantized Hall resistance of 3h/2e2 around ν = -2/3, respectively. Concomitantly, the longitudinal resistance exhibits distinct minima around ν = -1 and -2/3. The application of an electric field induces topological quantum phase transition from the IQAH state to a charge transfer insulator at ν = -1, and from the FQAH state to a generalized Wigner crystal state, further transitioning to a metallic state at ν = -2/3. Our study paves the way for the investigation of fractionally charged excitations and anyonic statistics at zero magnetic field based on semiconductor moiré materials.

cond-mat.mes-hall↗
arXiv2023

Rep2wav: Noise Robust text-to-speech Using self-supervised representations

Benefiting from the development of deep learning, text-to-speech (TTS) techniques using clean speech have achieved significant performance improvements. The data collected from real scenes often contains noise and generally needs to be denoised by speech enhancement models. Noise-robust TTS models are often trained using the enhanced speech, which thus suffer from speech distortion and background noise that affect the quality of the synthesized speech. Meanwhile, it was shown that self-supervised pre-trained models exhibit excellent noise robustness on many speech tasks, implying that the learned representation has a better tolerance for noise perturbations. In this work, we therefore explore pre-trained models to improve the noise robustness of TTS models. Based on HiFi-GAN, we first propose a representation-to-waveform vocoder, which aims to learn to map the representation of pre-trained models to the waveform. We then propose a text-to-representation FastSpeech2 model, which aims to learn to map text to pre-trained model representations. Experimental results on the LJSpeech and LibriTTS datasets show that our method outperforms those using speech enhancement methods in both subjective and objective metrics. Audio samples are available at: https://zqs01.github.io/rep2wav.

eess.AS↗
arXiv2023

Text Matching Improves Sequential Recommendation by Reducing Popularity Biases

This paper proposes Text mAtching based SequenTial rEcommendation model (TASTE), which maps items and users in an embedding space and recommends items by matching their text representations. TASTE verbalizes items and user-item interactions using identifiers and attributes of items. To better characterize user behaviors, TASTE additionally proposes an attention sparsity method, which enables TASTE to model longer user-item interactions by reducing the self-attention computations during encoding. Our experiments show that TASTE outperforms the state-of-the-art methods on widely used sequential recommendation datasets. TASTE alleviates the cold start problem by representing long-tail items using full-text modeling and bringing the benefits of pretrained language models to recommendation systems. Our further analyses illustrate that TASTE significantly improves the recommendation accuracy by reducing the popularity bias of previous item id based recommendation models and returning more appropriate and text-relevant items to satisfy users. All codes are available at https://github.com/OpenMatch/TASTE.

cs.IR↗
arXiv2023

Scaling Clinical Trial Matching Using Large Language Models: A Case Study in Oncology

Clinical trial matching is a key process in health delivery and discovery. In practice, it is plagued by overwhelming unstructured data and unscalable manual processing. In this paper, we conduct a systematic study on scaling clinical trial matching using large language models (LLMs), with oncology as the focus area. Our study is grounded in a clinical trial matching system currently in test deployment at a large U.S. health network. Initial findings are promising: out of box, cutting-edge LLMs, such as GPT-4, can already structure elaborate eligibility criteria of clinical trials and extract complex matching logic (e.g., nested AND/OR/NOT). While still far from perfect, LLMs substantially outperform prior strong baselines and may serve as a preliminary solution to help triage patient-trial candidates with humans in the loop. Our study also reveals a few significant growth areas for applying LLMs to end-to-end clinical trial matching, such as context limitation and accuracy, especially in structuring patient information from longitudinal medical records.

cs.CL↗
arXiv2023

Distilling Large Language Models for Biomedical Knowledge Extraction: A Case Study on Adverse Drug Events

Large language models (LLMs), such as GPT-4, have demonstrated remarkable capabilities across a wide range of tasks, including health applications. In this paper, we study how LLMs can be used to scale biomedical knowledge curation. We find that while LLMs already possess decent competency in structuring biomedical text, by distillation into a task-specific student model through self-supervised learning, substantial gains can be attained over out-of-box LLMs, with additional advantages such as cost, efficiency, and white-box model access. We conduct a case study on adverse drug event (ADE) extraction, which is an important area for improving care. On standard ADE extraction evaluation, a GPT-3.5 distilled PubMedBERT model attained comparable accuracy as supervised state-of-the-art models without using any labeled data. Despite being over 1,000 times smaller, the distilled model outperformed its teacher GPT-3.5 by over 6 absolute points in F1 and GPT-4 by over 5 absolute points. Ablation studies on distillation model choice (e.g., PubMedBERT vs BioGPT) and ADE extraction architecture shed light on best practice for biomedical knowledge extraction. Similar gains were attained by distillation for other standard biomedical knowledge extraction tasks such as gene-disease associations and protected health information, further illustrating the promise of this approach.

cs.CL↗
arXiv2023

Swarm of One: Bottom-up Emergence of Stable Robot Bodies from Identical Cells

Unlike most human-engineered systems, biological systems are emergent from low-level interactions, allowing much broader diversity and superior adaptation to the complex environments. Inspired by the process of morphogenesis in nature, a bottom-up design approach for robot morphology is proposed to treat a robot's body as an emergent response to underlying processes rather than a predefined shape. This paper presents Loopy, a "Swarm-of-One" polymorphic robot testbed that can be viewed simultaneously as a robotic swarm and a single robot. Loopy's shape is determined jointly by self-organization and morphological computing using physically linked homogeneous cells. Experimental results show that Loopy can form symmetric shapes consisting of lobes. Using the the same set of parameters, even small amounts of initial noise can change the number of lobes formed. However, once in a stable configuration, Loopy has an "inertia" to transfiguring in response to dynamic parameters. By making the connections among self-organization, morphological computing, and robot design, this paper lays the foundation for more adaptable robot designs in the future.

cs.RO↗
arXiv2023

Self-supervised Interest Point Detection and Description for Fisheye and Perspective Images

Keypoint detection and matching is a fundamental task in many computer vision problems, from shape reconstruction, to structure from motion, to AR/VR applications and robotics. It is a well-studied problem with remarkable successes such as SIFT, and more recent deep learning approaches. While great robustness is exhibited by these techniques with respect to noise, illumination variation, and rigid motion transformations, less attention has been placed on image distortion sensitivity. In this work, we focus on the case when this is caused by the geometry of the cameras used for image acquisition, and consider the keypoint detection and matching problem between the hybrid scenario of a fisheye and a projective image. We build on a state-of-the-art approach and derive a self-supervised procedure that enables training an interest point detector and descriptor network. We also collected two new datasets for additional training and testing in this unexplored scenario, and we demonstrate that current approaches are suboptimal because they are designed to work in traditional projective conditions, while the proposed approach turns out to be the most effective.

cs.CV↗
arXiv2023

Structure-Aware Language Model Pretraining Improves Dense Retrieval on Structured Data

This paper presents Structure Aware Dense Retrieval (SANTA) model, which encodes user queries and structured data in one universal embedding space for retrieving structured data. SANTA proposes two pretraining methods to make language models structure-aware and learn effective representations for structured data: 1) Structured Data Alignment, which utilizes the natural alignment relations between structured data and unstructured data for structure-aware pretraining. It contrastively trains language models to represent multi-modal text data and teaches models to distinguish matched structured data for unstructured texts. 2) Masked Entity Prediction, which designs an entity-oriented mask strategy and asks language models to fill in the masked entities. Our experiments show that SANTA achieves state-of-the-art on code search and product search and conducts convincing results in the zero-shot setting. SANTA learns tailored representations for multi-modal text data by aligning structured and unstructured data pairs and capturing structural semantics by masking and predicting entities in the structured data. All codes are available at https://github.com/OpenMatch/OpenMatch.

cs.IR↗
arXiv2023

ReSup: Reliable Label Noise Suppression for Facial Expression Recognition

Because of the ambiguous and subjective property of the facial expression recognition (FER) task, the label noise is widely existing in the FER dataset. For this problem, in the training phase, current FER methods often directly predict whether the label of the input image is noised or not, aiming to reduce the contribution of the noised data in training. However, we argue that this kind of method suffers from the low reliability of such noise data decision operation. It makes that some mistakenly abounded clean data are not utilized sufficiently and some mistakenly kept noised data disturbing the model learning process. In this paper, we propose a more reliable noise-label suppression method called ReSup (Reliable label noise Suppression for FER). First, instead of directly predicting noised or not, ReSup makes the noise data decision by modeling the distribution of noise and clean labels simultaneously according to the disagreement between the prediction and the target. Specifically, to achieve optimal distribution modeling, ReSup models the similarity distribution of all samples. To further enhance the reliability of our noise decision results, ReSup uses two networks to jointly achieve noise suppression. Specifically, ReSup utilize the property that two networks are less likely to make the same mistakes, making two networks swap decisions and tending to trust decisions with high agreement. Extensive experiments on three popular benchmarks show that the proposed method significantly outperforms state-of-the-art noisy label FER methods by 3.01% on FERPlus becnmarks. Code: https://github.com/purpleleaves007/FERDenoise

cs.CV↗
arXiv2023

Eeg2vec: Self-Supervised Electroencephalographic Representation Learning

Recently, many efforts have been made to explore how the brain processes speech using electroencephalographic (EEG) signals, where deep learning-based approaches were shown to be applicable in this field. In order to decode speech signals from EEG signals, linear networks, convolutional neural networks (CNN) and long short-term memory networks are often used in a supervised manner. Recording EEG-speech labeled data is rather time-consuming and laborious, while unlabeled EEG data is abundantly available. Whether self-supervised methods are helpful to learn EEG representation to boost the performance of EEG auditory-related tasks has not been well explored. In this work, we first propose a self-supervised model based on contrastive loss and reconstruction loss to learn EEG representations, and then use the obtained pre-trained model as a feature extractor for downstream tasks. Second, for two considered downstream tasks, we use CNNs and Transformer networks to learn local features and global features, respectively. Finally, the EEG data from other channels are mixed into the chosen EEG data for augmentation. The effectiveness of our method is verified on the EEG match-mismatch and EEG regression tasks of the ICASSP2023 Auditory EEG Challenge.

eess.AS↗
arXiv2023

Wital: A COTS WiFi Devices Based Vital Signs Monitoring System Using NLOS Sensing Model

Vital sign (breathing and heartbeat) monitoring is essential for patient care and sleep disease prevention. Most current solutions are based on wearable sensors or cameras; however, the former could affect sleep quality, while the latter often present privacy concerns. To address these shortcomings, we propose Wital, a contactless vital sign monitoring system based on low-cost and widespread commercial off-the-shelf (COTS) Wi-Fi devices. There are two challenges that need to be overcome. First, the torso deformations caused by breathing/heartbeats are weak. How can such deformations be effectively captured? Second, movements such as turning over affect the accuracy of vital sign monitoring. How can such detrimental effects be avoided? For the former, we propose a non-line-of-sight (NLOS) sensing model for modeling the relationship between the energy ratio of line-of-sight (LOS) to NLOS signals and the vital sign monitoring capability using Ricean K theory and use this model to guide the system construction to better capture the deformations caused by breathing/heartbeats. For the latter, we propose a motion segmentation method based on motion regularity detection that accurately distinguishes respiration from other motions, and we remove periods that include movements such as turning over to eliminate detrimental effects. We have implemented and validated Wital on low-cost COTS devices. The experimental results demonstrate the effectiveness of Wital in monitoring vital signs.

cs.HC↗
arXiv2023

Integrated Artificial Neural Network with Trainable Activation Function Enabled by Topological Insulator-based Spin-Orbit Torque Devices

Non-volatile memristors offer a salient platform for artificial neural network (ANN), but the integration of different function blocks into one hardware system remains challenging. Here we demonstrate the implementation of brain-like synaptic (SOT-S) and neuronal (SOT-N) functions in the Bi2Te3/CrTe2 heterostructure-based spin-orbit torque (SOT) device. The SOT-S unit exhibits highly linear (linearity error < 4.19%) and symmetrical long-term potentiation/depression process, resulting in better performance compared to other memristor synapses. Meanwhile, the Sigmoid-shape transition curve inherited in the SOT-N cell replaces the software-based activation function block, hence reducing the system complexity. On this basis, we employ a serial-connected, voltage-mode sensing ANN architecture to enhance the vector-matrix multiplication signal strength with low reading error of 0.61%. Furthermore, the trainable activation function of SOT-N enables the integrated SOT-ANN to execute the Batch Normalization algorithm and activation operation within one clock cycle, which bring about improved on/off-chip training performance close to the ideal baseline.

cond-mat.mes-hall↗
arXiv2023

Rhino: An Autonomous Robot for Mapping Underground Mine Environments

There are many benefits for exploring and exploiting underground mines, but there are also significant risks and challenges. One such risk is the potential for accidents caused by the collapse of the pillars, and roofs which can be mitigated through inspections. However, these inspections can be costly and may put the safety of the inspectors at risk. To address this issue, this work presents Rhino, an autonomous robot that can navigate underground mine environments and generate 3D maps. These generated maps will allow mine workers to proactively respond to potential hazards and prevent accidents. The system being developed is a skid-steer, four-wheeled unmanned ground vehicle (UGV) that uses a LiDAR and IMU to perform long-duration autonomous navigation and generation of maps through a LIO-SAM framework. The system has been tested in different environments and terrains to ensure its robustness and ability to operate for extended periods of time while also generating 3D maps.

cs.RO↗