Is this you? Claim this profile to correct it, add a bio and choose the work people see first.
Claim this profileWorks92 from public data
- EnlightenGAN: Deep Light Enhancement Without Paired Supervision2,715
This paper proposes a highly effective unsupervised generative adversarial network, dubbed EnlightenGAN, that can be trained without low/normal-light image pairs, yet proves to generalize very well on various real-world test images.
- Patient Knowledge Distillation for BERT Model Compression1,021
This work proposes a Patient Knowledge Distillation approach to compress an original large model (teacher) into an equally-effective lightweight shallow network (student), which translates into improved results on multiple NLP tasks with a significant gain in training efficiency, without sacrificing model accuracy.
- 812
- 811
- 761
- Risk Prediction with Electronic Health Records: A Deep Learning Approach433
A deep learning approach for phenotyping from patient EHRs by building a four-layer convolutional neural network model for extracting phenotypes and perform prediction and the proposed model is validated on a real world EHR data warehouse under the scenario of predictive modeling of chronic diseases.
- Fully-Adaptive Feature Sharing in Multi-Task Networks with Applications in Person Attribute Classification410
Evaluation on person attributes classification tasks involving facial and clothing attributes suggests that the models produced by the proposed method are fast, compact and can closely match or exceed the state-of-the-art accuracy from strong baselines by much more expensive models.
- 384
- Jointly Attentive Spatial-Temporal Pooling Networks for Video-Based Person Re-identification341
This work presents a novel joint Spatial and Temporal Attention Pooling Network (ASTPN) for video-based person re-identification, which enables the feature extractor to be aware of the current input video sequences, in a way that interdependency from the matching items can directly influence the computation of each other's representation.
- 341
- 268
- 234
- Diverse Few-Shot Text Classification with Multiple Metrics228
This work proposes an adaptive metric learning approach that automatically determines the best weighted combination from a set of metrics obtained from meta-training tasks for a newly seen few-shot task.
- 195
- Merge, Then Compress: Demystify Efficient SMoE with Hints from Its Routing Policy139
M-SMoE is proposed, which leverages routing statistics to guide expert merging and further decomposes the merged experts into low-rank and structural sparse alternatives, thus diminishing the impact of insignificant experts.
- Multi-Fact Correction in Abstractive Text Summarization135
Span-Fact is a suite of two factual correction models that leverages knowledge learned from question answering models to make corrections in system-generated summaries via span selection to significantly boost the factual consistency of system- generated summaries without sacrificing summary quality in terms of both automatic metrics and human evaluation.
- Measuring Patient Similarities via a Deep Architecture with Medical Concept Embedding122
A patient similarity evaluation framework based on temporal matching of longitudinal patient EHRs, which takes a convolutional neural network architecture, and learns an optimal representation of patient clinical record through medical concept embedding.
- Measuring Patient Similarities via a Deep Architecture with Medical Concept Embedding122
A patient similarity evaluation framework based on temporal matching of longitudinal patient EHRs, which takes a convolutional neural network architecture, and learns an optimal representation of patient clinical record through medical concept embedding.
- 120
- Walk and Learn: Facial Attribute Representation Learning from Egocentric Video and Contextual Data116
This work tracks the faces of casual walkers on more than 40 hours of egocentric video and automatically extracts nearly 5 million pairs of images connected by or from different face tracks, along with their weather and location context, under pose and lighting variations, to learn a rich feature representation for facial attribute classification.
- EarlyBERT: Efficient BERT Training via Early-bird Lottery Tickets113
By slimming the self-attention and fully-connected sub-layers inside a transformer, this work is the first to identify structured winning tickets in the early stage of BERT training.
- 90
- Contrastive Distillation on Intermediate Representations for Language Model Compression87
CoDIR is proposed, a principled knowledge distillation framework where the student is trained to distill knowledge through intermediate layers of the teacher via a contrastive objective, and achieves superb performance on the GLUE benchmark, outperforming state-of-the-art compression methods.
- Mining millions of reviews85
A product ranking model that applies weights to product review factors to calculate a products ranking score is presented which produces product ranking results that closely relate to the products sales ranking as reported by the retailer.
- 77
- Segment convolutional neural networks (Seg-CNNs) for classifying relations in clinical notes74
The proposed Segment Convolutional Neural Networks (Seg-CNNs) show state-of-the-art performance while requiring no manual feature engineering, and it is shown that medical domain word embeddings help improve relation classification.
- A hybrid approach with optimization-based and metric-based meta-learner for few-shot learning66
This work proposes a hybrid meta-learning model called Meta-Metric-Learner which combines the merits of both optimization- and metric-based approaches and is able to handle flexible numbers of classes as well as generate more generalized metrics for classification across tasks.
- 60
- 60
- Attend to count: Crowd counting with adaptive capacity multi-scale CNNs57
The Adaptive Capacity Multi-scale convolutional neural networks (ACM-CNN), a novel crowd counting approach which can assign different capacities to different portions of the input, is proposed.
- 57
- Fast Neural Networks with Circulant Projections.50
This work proposes to replace the conventional linear projection with the circulant projection, which enables the use of the Fast Fourier Transform to speed up the computation of a fully-connected neural network layer.
- Not All Inputs Are Valid: Towards Open-Set Video Moment Retrieval using Language49
This paper proposes a novel model OpenVMR, which first distinguishes ID and OOD queries based on the normalizing flow technology, and then conducts moment retrieval based on ID queries, and first distinguishes ID and OOD queries based on the normalizing flow technology, and then conducts moment retrieval based on ID queries.
- Efficient Robust Training via Backward Smoothing46
This work develops a new understanding towards Fast Adversarial Training, by viewing random initialization as performing randomized smoothing for better optimization of the inner maximization problem, and proposes a new initialization strategy, backward smoothing, to further improve the stability and model robustness over single-step robust training methods.
- A probabilistic disease progression modeling approach and its application to integrated Huntington’s disease observational data43
A framework to build probabilistic disease progression models using observational medical data that addresses practical challenges in observational data and can help enhance the understanding of progression of chronic diseases.
- Rethinking Weakly-Supervised Video Temporal Grounding From a Game Perspective42
This is the first attempt to tackle the challenging task of weakly-supervised video temporal grounding from a novel game perspective, which effectively learns the uncertain relationship between each vision-language pair with diverse granularity and flexible combination for multi-level cross-modal interaction.
- 38
- APo-VAE: Text Generation in Hyperbolic Space36
Extensive experiments in language modeling, unaligned style transfer, and dialog-response generation demonstrate the effectiveness of the proposed APo-VAE model over VAEs in Euclidean latent space, thanks to its superb capabilities in capturing latent language hierarchies in hyperbolic space.
- Cluster-Former: Clustering-based Sparse Transformer for Question Answering35
Cluster-Former is proposed, a novel clustering-based sparse Transformer to perform attention across chunked sequences that allows information integration beyond local windows, which is especially beneficial for question answering (QA) tasks that rely on long-range dependencies.
- Mixed-Supervised Dual-Network for Medical Image Segmentation29
This paper proposes Mixed-Supervised Dual-Network (MSDN), a novel architecture which consists of two separate networks for the detection and segmentation tasks respectively, and a series of connection modules between the layers of the two networks.
- 26
- Robustness Challenges in Model Distillation and Pruning for Natural Language Understanding25
This work studies two popular model compression techniques including knowledge distillation and pruning and shows that the compressed models are significantly less robust than their PLM counterparts on OOD test sets although they obtain similar performance on in-distribution development sets for a task.
- 24
- 23
- A Quasi-ARX Neural Network with Switching Mechanism to Adaptive Control of Nonlinear Systems20
An improved quasi-ARX neural network is introduced and its application to adaptive control of nonlinear systems is discussed and some stability analysis of the control system is shown.
- Deep Multimodality Model for Multi-task Multi-view Learning18
A deep multi-task multi-view learning framework that learns a deep representation for such dual-heterogeneity problems and empirical studies on multiple real-world data sets demonstrate the effectiveness of the proposed Deep-MTMV algorithm.
- 18
- 17
- Each Rank Could be an Expert: Single-Ranked Mixture of Experts LoRA for Multi-task Learning16
This paper demonstrates that the dynamic routing of multiple LoRAs is functionally equivalent to rank partitioning and block-level activation within a single LoRA, and proposes Single-ranked Mixture of Experts LoRA, which embeds MoE into LoRA by treating each rank as an independent expert.
- 16
- From Head to Tail: Towards Balanced Representation in Large Vision-Language Models through Adaptive Data Calibration13
An in-depth analysis of the LT issues in LVLMs is conducted and an Adaptive Data Refinement Framework (ADR) is proposed, which effectively mitigates the long-tail problem in the training data, improving the average performance of LLaVA 1.5 relatively by 4.36%, without increasing the training data volume.
- 13
- RE-TRAC: REcursive TRAjectory Compression for Deep Search Agents12
Re-TRAC is proposed, an agentic framework that performs cross-trajectory exploration by generating a structured state representation after each trajectory to summarize evidence, uncertainties, failures, and future plans, and conditioning subsequent trajectories on this state representation, enabling iterative reflection and globally informed planning.
- A novel dual-network architecture for mixed-supervised medical image segmentation12
A novel network architecture called Mixed-Supervised Dual-Network (MSDN), which consists of two separate networks for the segmentation and detection tasks respectively, and a series of connection modules between the layers of the two networks, which is more effective and stable.
- 12
- 12
- 12
- 11
- HGM-cNet: Integrating hippocampal gray matter probability map into a cascaded deep learning framework improves hippocampus segmentation11
The integration of the HGM probability map in the cascaded deep learning framework was demonstrated to facilitate capturing hippocampal atrophy more accurately than alternative methods in AD analysis and achieved a superior hippocampus segmentation performance in each group of cognitive normal, mild cognitive impairment, and AD.
- 11
- An Exploration of Latent Structure in Observational Huntington's Disease Studies.10
Hidden low dimensional structure in the data that correlates well with surrogate measures of HD progression are discovered and are promising candidates for future consumption by downstream statistical HD progression models.
- 10
- 9
- 9
- Sentiment identification by incorporating syntax, semantics and context information8
A method based on conditional random fields to incorporate sentence structure (syntax and semantics) and context information to identify sentiments of sentences within a document is proposed.
- 7
- Crowdsourcing recommendations from social sentiment6
This paper investigates an innovative recommendation system by incorporating relevant social opinion and sentiment information by incorporating user-generated high-level semantic sentiment in the analysis and applies its models to real time data and observes promising results for not only product recommendation but also job recommendation.
- Design and Optimization of Big Data and Machine Learning-Based Risk Monitoring System in Financial Markets5
Research findings demonstrate that the system significantly enhances efficiency and accuracy in risk management, particularly excelling in identifying and warning against market crash risks.
- 5
- 4
- 4
- Switching-State Dynamical Modeling of Daily Behavioral Data4
This work focuses on modeling heterogeneous dynamics to better capture health status changes under different conditions, which may lead to more effective state-dependent intervention strategies and implements switching-state dynamic system models with different complexity levels on real-world daily behavioral data.
- A Data-Driven Method for Generating Robust Symptom Onset Indicators in Huntington's Disease Registry Data.4
A disease registry can be a powerful tool to observe the course of disease, to understand variations in treatment and outcomes, to examine factors that influence prognosis and quality of life, and to assess the clinical cost andquality of care.
- 4
- 4
- Bayesian CycleGAN via Marginalizing Latent Sampling.3
The proposed Bayesian CycleGAN explores the full posteriors of Bayesian cyclic model and optimizes the model with maximum a posteriori (MAP) estimation and can achieve more stable training, superior performance and diversified images generating.
- Legislative prediction with dual uncertainty minimization from heterogeneous information3
A novel prediction model is presented that maximizes the usage of publicly accessible heterogeneous data, i.e., bill text and lawmakers' profile data, to carry out effective legislative prediction and provides superior prediction accuracy with visible performance gain.
- 3
- 2
- 2
- 2
- 2
- 2
- 2
- 2
- The integration of traditional culture and educational drama for young children based on the AdaBoost model1
The results show that the AdaBoost-WBLS model has significantly improved G-mean on dichotomous and multiclassified datasets compared with similar models, with higher accuracy and better ability to handle unbalanced data.
- 1
- IBM-Northwestern@Trecvid 2014: Surveillance event detection1
An interactive approach to visualize data with temporal relations and a novel risk ranking method to differentiate detection results and present more informative ones to the user for better interaction are proposed.
- –
- –
- [Lossless compression of high sampling rate ECG data based on BW algorithm].–
A BW-based high sampling rate ECG data lossless compression algorithm is proposed, which has gained much improvement in terms of the compression ratio, especially when applied to the high sampling rates ECGData.
- Design and Implementation of an Intelligent Decision Support System–
This paper redesigns the IDSS by applying object-oriented technology, and improves its performance.
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.