Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks150 from public data
- Earthformer: Exploring Space-Time Transformers for Earth System Forecasting416
The idea is to decompose the data into cuboids and apply cuboid-level self-attention in parallel, which achieves state-of-the-art performance on real-world benchmarks about precipitation nowcasting and El Nino/Southern Oscillation forecasting.
- 365
- AIM: Adapting Image Models for Efficient Video Action Recognition273
This work proposes a novel method to Adapt pre-trained Image Models (AIM) for efficient video understanding, and shows that the proposed AIM can achieve competitive or even better performance than prior arts with substantially fewer tunable parameters on four video action recognition benchmarks.
- Towards Geospatial Foundation Models via Continual Pretraining169
This work investigates the potential of continual pretraining from large-scale ImageNet-22k models and proposes a multi-objective continual pretraining paradigm, which leverages the strong representations of ImageNet while simultaneously providing the freedom to learn valuable in-domain features.
- Visual Prompt Tuning for Test-time Domain Adaptation124
This work tackles the practical yet challenging test-time adaptation (TTA) problem, where a model adapts to the target domain without accessing the source data, with a simple recipe called Data-efficient Prompt Tuning (DePT).
- Can We Use Arbitrary Objects to Attack LiDAR Perception in Autonomous Driving?102
This paper investigates an easier way to perform effective adversarial attacks with high flexibility and good stealthiness against LiDAR perception in autonomous driving and proposes a novel attack framework based on which the attacker can identify a few adversarial locations in the physical space.
- 84
- Parsing Tweets into Universal Dependencies77
It is shown that it is challenging to deliver consistent annotation due to ambiguity in understanding and explaining tweets and proposed a new method to distill an ensemble of 20 transition-based parsers into a single one that achieves an improvement of 2.2 in LAS over the un-ensembled baseline and outperforms parsers that are state-of-the-art on other treebanks in both accuracy and speed.
- 75
- Adversarial Attacks against LiDAR Semantic Segmentation in Autonomous Driving73
This paper proposes a novel adversarial attack framework based on which the attacker can easily fool LiDAR semantic segmentation by placing some simple objects at some locations in the physical space, and conducts extensive real-world experiments to evaluate the performance of the proposed attack framework.
- Partial and Asymmetric Contrastive Learning for Out-of-Distribution Detection in Long-Tailed Recognition70
This work demonstrates that existing OOD detection methods commonly suffer from significant performance degradation when the training set is long-tail distributed, and proposes Partial and Asymmetric Supervised Contrastive Learning (PASCL), which explicitly encourages the model to distinguish between tail-class in-distribution samples and OOD samples.
- A Closer Look at Few-Shot Crosslingual Transfer: The Choice of Shots Matters70
This work conducts a large-scale experimental study on 40 sets of sampled few shots for six diverse NLP tasks across up to 40 languages, and shows that a straightforward full model finetuning approach is quite effective for few-shot transfer, outperforming several state-of-the-art few- shot approaches.
- Unsupervised Semantic Segmentation with Self-supervised Object-centric Representations63
A methodology based on unsupervised saliency masks and self-supervised feature clustering to kickstart object discovery followed by training a semantic segmentation network on pseudo-labels to bootstrap the system on images with multiple objects is proposed.
- ViewCo: Discovering Text-Supervised Segmentation Masks via Multi-View Semantic Consistency62
The text-to-views consistency modeling benefits the dense assignment of the visual features by encouraging different crops to align with the same text, while the cross-view segmentation consistency modeling provides additional self-supervision, overcoming the limitation of ambiguous text supervision for segmentation masks.
- Land use classification using convolutional neural networks applied to ground-level images53
This work presents a general framework that uses ground-level images from Flickr for land use mapping, and utilizes high-level semantic image features extracted using deep learning, specifically convolutional neural networks, which allow it to achieve upwards of 76% accuracy on a challenging eight class land use maps problem.
- 50
- 48
- CoupAlign: Coupling Word-Pixel with Sentence-Mask Alignments for Referring Image Segmentation48
CoupAlign is proposed, a simple yet effective multi-level visual-semantic alignment method to couple sentence-mask alignment with word-pixel alignment to enforce object mask constraint for achieving more accurate localization and segmentation.
- Towards Backdoor Attacks against LiDAR Object Detection in Autonomous Driving47
A novel backdoor attack strategy is proposed based on which the attacker can achieve the attack goal by poisoning a small number of point cloud samples and it is physically realizable, and it allows the attacker to easily perform the attack using some common objects as the triggers.
- MetaWave: Attacking mmWave Sensing with Meta-material-enhanced Tags39
This paper proposes and design a novel passive mmWave attack with low-cost and easily obtainable meta-material tags for both vanish and ghost attack types, and demonstrates that tags made of ordinal material can be leveraged to precisely tamper the mmWave echo signal and spoof the range, angle, and speed sensing measurements.
- Malicious Attacks against Multi-Sensor Fusion in Autonomous Driving38
The first study on the vulnerability of multi-sensor fusion systems that employ LiDAR, camera, and radar is presented, which proposes a novel attack method that can simultaneously attack all three types of sensing modalities using a single type of adversarial object.
- Fine-Grained Visual–Text Prompt-Driven Self-Training for Open-Vocabulary Object Detection36
This article proposes a simple but effective fine-grained visual-text prompt-driven self-training paradigm for open-vocabulary detection (VTP-OVD) that introduces a fine-grained visual–text prompt adapting stage to enhance the current self-training paradigm with a more powerful fine-grained alignment.
- A Systematic Study of Leveraging Subword Information for Learning Word Representations35
This work proposes a general framework for learning subword-informed word representations that allows for easy experimentation with different segmentation and composition components, also including more advanced techniques based on position embeddings and self-attention.
- Contrastive Instruction-Trajectory Learning for Vision-Language Navigation33
A Contrastive Instruction-Trajectory Learning (CITL) framework that explores invariance across similar data samples and variance across different ones to learn distinctive representations for robust navigation and can be easily integrated with VLN backbones to form a new learning paradigm and achieve better generalizability in unseen environments.
- 29
- Sentence Simplification via Large Language Models27
Experimental results show LLMs outperform state-of-the-art sentence simplification methods, and are judged to be on a par with human annotators.
- Development of a position–velocity–time-modulated two-dimensional ion beam figuring system for synchrotron x-ray mirror fabrication27
The Optical Metrology Group at National Synchrotron Light Source II has designed and built a position-velocity-time-modulated two-dimensional IBF system (PVT-IBF) with three novel characteristics: a beam footprint on the mirror was used as a reference to align the coordinate systems between the metrology and the IBF hardware and the robust iterative Fourier transform-based dwell time algorithm was applied to obtain an accurate dwell time map.
- MixReorg: Cross-Modal Mixed Patch Reorganization is a Good Mask Learner for Open-World Semantic Segmentation25
MixReorg is proposed, a novel and straightforward pre-training paradigm for semantic segmentation that enhances a model’s ability to reorganize patches mixed across images, exploring both local visual relevance and global semantic coherence.
- TileMask: A Passive-Reflection-based Attack against mmWave Radar Object Detection in Autonomous Driving25
This paper investigates the possibility of using a few adversarial objects to attack the DNN-based radar object detection models through passive reflection using low-cost, readily-available and easily concealable geometric shaped objects.
- On the Importance of Subword Information for Morphological Tasks in Truly Low-Resource Languages25
A comprehensive analysis focused on the usefulness of subwords for word representation learning in truly low-resource scenarios and for three representative morphological tasks: fine-grained entity typing, morphological tagging, and named entity recognition shows that subword-informed models are universally useful across all language types.
- 24
- 24
- Towards Medical Artificial General Intelligence via Knowledge-Enhanced Multimodal Pretraining21
The proposed MOTOR successfully mimics the human practice of fulfilling a"medical student" to accelerate the process of becoming a"specialist" and believes that this work makes a significant stride in realizing MAGI.
- On Constructing Confidence Region for Model Parameters in Stochastic Gradient Descent Via Batch Means20
An easy-to-implement algorithm to construct asymptotically valid confidence regions for model parameters in stochastic gradient descent is studied and a process-level functional central limit theorem for Polyak-Ruppert averaging iterates is established.
- Benchmarking Zero-Shot Recognition with Vision-Language Models: Challenges on Granularity and Specificity19
Novel benchmarks for evaluating vision-language models in zero-shot recognition show that VLMs favor moderately fine-grained concepts and struggle with specificity, often misjudging texts that differ from their training data.
- Language-Driven Visual Consensus for Zero-Shot Semantic Segmentation18
Equipped with a vision-language prompting strategy, this approach significantly boosts the generalization capacity of segmentation models for unseen classes and introduces route attention into the transformer decoder to find visual consensus, thereby enhancing semantic consistency within the same object.
- Improving adversarial robustness of deep neural networks by using semantic information18
Experimental results on the MNIST and CIFAR-10 datasets show that this approach greatly improves adversarial robustness even using a very small dataset from the training data; moreover, it can defend against FGSM adversarial attacks that have a completely different pattern from the model seen during retraining.
- 18
- Benchmarking Robustness of Multimodal Image-Text Models under Distribution Shift17
It is observed that multimodal models are not robust to image and text perturbations, especially to image perturbations, so several new multimodal robustness benchmarks are proposed by applying 17 image perturbation and 16 text perturbation techniques on top of existing datasets.
- Chinese Idiom Paraphrasing16
A novel task to rephrase idiom-containing sentences to non-idiomatic ones under the premise of preserving the original sentence’s meaning is proposed, and a novel infill-based approach based on text infilling is proposed that has better performance than the baselines based on the established CIP dataset.
- Optical transient grating pumped X-ray diffraction microscopy for studying mesoscale structural dynamics16
TGXD successfully characterizes mesoscopic energy transport in functional materials without relying on a specific transport model and provides spatial resolution within the engineered optical excitation profile.
- 14
- 12
- 12
- A Simple BERT-Based Approach for Lexical Simplification12
This work presents a simple BERT-based LS approach that makes use of the pre-trained unsupervised deep bidirectional representations BERT, and experimental results show that this approach obtains obvious improvement than these baselines leveraging linguistic databases and parallel corpus.
- 11
- Joint user profiling with hierarchical attention networks11
This paper proposes a joint user profiling model with hierarchical attention networks (JUHA) to learn informative user representations for user profiling and demonstrates that this approach outperforms several state-of-the-art methods and improves prediction performance.
- Gated Multimodal Fusion with Contrastive Learning for Turn-Taking Prediction in Human-Robot Dialogue11
- Study on the Multi-Equipment Integrated Scheduling Problem of a U-Shaped Automated Container Terminal Based on Graph Neural Network and Deep Reinforcement Learning10
This study investigates the integrated scheduling problem of DTQCs, IGVs and DCRCs under the hybrid “loading and unloading” mode, expresses the problem as a Markovian decision-making process, and establishes a disjunctive graph model.
- SuperScaler: Supporting Flexible DNN Parallelization via a Unified Abstraction10
SuperScaler is a system that facilitates the design and generation of highly flexible parallelization plans that formulates the plan design andgeneration into three sequential phases explicitly: model transformation, space-time scheduling, and data dependency preserving.
- 10
- 10
- 10
- 9
- An Online Defense against Object-based LiDAR Attacks in Autonomous Driving9
This work proposes a novel online defense mechanism against object-based LiDAR attacks, aiming to identify and remove the adversarial LiDAR points generated by the objects used by attackers before the data is fed into the perception module of autonomous driving systems.
- 8
- Driver Behavior-aware Parking Availability Crowdsensing System Using Truth Discovery8
A crowdsensing system called SpotE that can provide spot-level availability in a parking lot using drivers’ smartphone sensors that avoids the high cost of installing additional sensors and enables large-scale outdoor deployment.
- 7
- 7
- Investigating Biases in COVID-19 Diagnostic Systems Processed with Automated Speech Anonymization Algorithms7
It is shown that although data augmentation helps to recover some loss in overall performance, it can lead to a larger discrepancy in performance for over-represented and under-represented groups.
- Evaluating the Utility of Five Gene Fragments for Genetic Diversity Analyses of Mytella strigata Populations7
It is indicated that COI is the most useful gene fragment for genetic diversity studies of M. strigata populations; D1 28S and 18S-ITS1 sequences would be useful for species identification because of their low intraspecific diversity.
- Bayesian Learning for Neural Dependency Parsing7
This work demonstrates that in the small data regime, where uncertainty around parameter estimation and model prediction matters the most, Bayesian neural modeling is very effective, and utilizes an efficient sampling procedure via stochastic gradient Langevin dynamics to generate samples from the approximated posterior.
- SparseGrasp: Robotic Grasping via 3D Semantic Gaussian Splatting from Sparse Multi-View RGB Images6
Experimental results show that SparseGrasp significantly outperforms state-of-the-art methods in terms of both speed and adaptability, providing a robust solution for multi-turn grasping in changeable environment.
- Pixel-level Correspondence for Self-Supervised Learning from Video6
Pixel-level Correspondence (PiCo) is proposed, a method for dense contrastive learning from video which outperforms self-supervised baselines on multiple dense prediction tasks, without compromising performance on image classification.
- Combining Deep Generative Models and Multi-lingual Pretraining for Semi-supervised Document Classification6
This work combines semi-supervised deep generative models and multi-lingual pretraining techniques to form a pipeline for document classification task and outperforms the state-of-the-art counterparts in low-resource settings across several languages.
- 6
- Asymmetry Vulnerability and Physical Attacks on Online Map Construction for Autonomous Driving5
This study presents the first vulnerability assessment of online map construction models and introduces the first digital and physical attack against them, which proposes asymmetric data fine-tuning as a targeted defense, which significantly improves model robustness.
- Multi-planar fixation of displaced, multi-fragmentary patella fractures in elderly patients with anterior locking plates and cerclage wires: preliminary results5
Anterior locking plates and cerclage wires in tandem provide reliable multi-planar fixation for displaced, multi-fragmentary patellar fractures in elderly patients and resulted in favorable clinical outcomes.
- SPT: Semi-Parametric Prompt Tuning for Multitask Prompted Learning5
SPT, a semi-parametric prompt tuning method for multitask prompted learning, which is a memory bank from where memory prompts are retrieved based on discrete prompts, is proposed.
- Cross-modal Transfer Learning via Multi-grained Alignment for End-to-End Spoken Language Understanding5
This work proposes a novel multi-grained cross-modal transfer learning framework for E2E-SLU with a cross attention module to align the tokens of text with the frame features of speech, encouraging the model to target at the salient acoustic features attended to each token during transferring the semantic information.
- 5
- 4
- Towards Real-Time Defense against Object-Based LiDAR Attacks in Autonomous Driving4
This work proposes the first real-time defense mechanism against object-based LiDAR attacks in autonomous driving, both detection model-agnostic and attack-agnostic, requiring no prior knowledge of the number, shape, size, or placement of adversarial objects.
- Dynamic Defense for Car-Borne LiDAR Vehicle Detection4
Evaluation based on open data and real-world experiments with embedded system implementation show that, when confronting adaptive attacks, Hyper3Def outperforms various baseline defenses including the adversarial training, which is often cited as the state of the art.
- 4
- SimCon Loss with Multiple Views for Text Supervised Semantic Segmentation4
A novel loss function termed SimCon, which accounts for intra-modal similarities to determine the appropriate set of positive samples to align, is proposed, which leads to robust training and faster convergence on zero-shot, text supervised semantic segmentation.
- 4
- 4
- Edge Deep Learning Enabled Freezing of Gait Detection in Parkinson's Patients4
The design of a wireless sensor network for detecting and alerting the freezing of gait symptoms in patients with Parkinson's disease is presented, which is self-contained and processes all user data locally without streaming data to external devices or the cloud, thus eliminating the cybersecurity risks and power penalty associated with wireless data transmission.
- 3
- 3
- 3
- 3
- Evaluation System of Foreign Language Teaching Quality Based on Spatiotemporal Feature Fusion3
A framework-based spatiotemporal modeling method is presented in this article that has higher accuracy and can be applied to the evaluation of foreign language teaching quality.
- 3
- 3
- 3
- 3
- 3
- 2
- AI4Reading: Chinese Audiobook Interpretation System Based on Multi-Agent Collaboration2
Compared with expert interpretations with the proposed AI4Reading system's output, the results show that although AI4Reading still has a gap in speech generation quality, the generated interpretative scripts are simpler and more accurate.
- 2
- 2
- 2
- 2
- Statistical Inference for Model Parameters in Stochastic Gradient Descent via Batch Means.2
A simple algorithm to construct asymptotically valid confidence regions for model parameters using the batch means method to cancel out the covariance matrix which is hard/costly to estimate.
- An Orderly Charging and Discharging Scheduling Strategy of Electric Vehicles Considering Demand Responsiveness2
Simulation results show that the strategy proposed can consider the influence of various indexes of EVs on schedule, and effectively realize the dispatching plan for aggregators.
- 2
- 2
- 1
- 1
- Demo: Attacking LiDAR Semantic Segmentation in Autonomous Driving1
It is demonstrated that existing LiDAR semantic segmentation models in autonomous driving systems can be easily fooled by placing some simple objects on the road, such as cardboard and traffic signs, which can hide a vehicle and change the road surface to road-side vegetation.
- 1
- 1
- 1
- Problems and Countermeasures of Educational Management of Postgraduates in Colleges and Universities1
- 1
- Research on the application of GIS and sensor network in the building settlement observation1
This paper sets up the model of the building's settlement observation and provides the property of the application about GIS technology and Sensor network and explores the process of setting up the real-time monitoring system who bases on sensor network technology for the data collection.
- 1
- Automatic Chinese Summarization Method Based on the HowNet and Clustering Algorithm1
To solve the problems in traditional automatic Chinese summarization, a new method based on the word concept and clustering is presented, which is more efficient and robust than the traditional one.
- The study of Xindice in the application of EMR1
This thesis presents a better plan to storage and query EMR based on XML, which considers that Xindice, as a representative Open Source information server which was specially signed for XML databases, is a good choice for constructing the storage hierarchy of XML system architecture.
- –
- –
- Deep Learning-Based Automated Kidney and Cortex Segmentation from Non-contrast T1-weighted Images–
This work develops a fully automated deep learning-based segmentation method for segmenting the entire kidney and internal structures in MR images, trained with non-contrast-enhanced T1-weight MR images from 40 volunteers, validated against manual segmentation.
- –
- –
- –
- –
- The design of low-power demodulation circuit for AmBC based on DTMB–
This paper provides a design scheme of a low-power demodulation circuit, in which the receiver can identify the frame header and frame body of the DTMB signal, select the corresponding comparison threshold and get the digital demodulated signal.
- –
- –
- A fast TDOA and VDOA estimation method of the wideband frequency hopping signal–
A fast and high-precision estimation algorithm for the time difference of arrival and velocity difference of wideband frequency hopping signals is proposed and the simulation experiment of the proposed method verifies the correctness and performance superiority of the algorithm.
- –
- –
- Face Concerns, Self-Stigma, Linguistic Factors, and Help-Seeking Intentions for Mental Health--Model–
- Interface design of three-dimensional virtual teaching situation based on psychological schema theory–
An interface design method suitable for 3D virtual teaching software is constructed, which can enhance the fluency in the operation process, reduce the psychological burden of students in the knowledge learning process, and provide services for better virtual simulation teaching experience.
- Statistics Based Neural Networks Method for Industrial Image Inspection–
A new method is presented by combining a statistics method with a neural networks method, which could reduce the interference of machine dynamics and improve the inspection accuracy, which has a better performance in comparison with traditional statistics approach.
- –
- –
- –
- –
- Extracting masts of overhead supply and street lights from point cloud–
Regular inspection and documentation for railway assets are necessary to monitor the status of the traffic environment and mobile Laser Scanning (MLS) makes it possible to collect highly accurate spatially accurate data.
- Research on the Ambiguity in Design Sketch and Variability in 3D Modeling in Automobile Design–
The research found that a certain range of ambiguity in design sketch and variability in 3D materialized processing based on 3D modeling parallelism show a positive effect on the possibility of design solution space.
- –
- –
- A rapid parameter selection method of kernel function–
A rapid hybrid parameter selection method of kernel function based on uniform design and grid search is proposed and effectiveness and generality of the method are proved by experiment.
- –
- –
- An improved scheduling algorithm for 3D cluster rendering with platform LSF–
A dynamic feedback load balance (DFLB) algorithm, the work principle of load sharing facility (LSF) and optimization of external scheduler plug-in are focused on.
- –
- Contact sheet system in the Department of Pharmacy–
Pharmaceutical sector in medical institutions should adopt a system of physician contact and contact sheet system served as a pharmacist review of evidence from work, and unreasonable interference in real-time prescription records, highlighting the value of hospital pharmacy.
- –
- –
- –
- –
- –
- –
- A Design of a Simple High Performance Microprocessor–
The instruction dependency checking algorithm and the data dependency Checking algorithm are brought up in this paper to settle the conflict of the correlative data happening in the executing of instruction.
- –
- –
Publication data from OpenAlex, with missing venues and authors filled in from Crossref; citation counts are the higher of OpenAlex and Semantic Scholar, last synced 2026-10-15. One-sentence summaries under some papers are written by Semantic Scholar’s model. Citation counts may be lower than on Google Scholar, which indexes more sources.
Report an error
Wrong papers, two people merged into one, or a profile that should not be here? Tell us and we will fix or hide it.