Hanwang Zhang lists you as their PhD student. Claim this profile to confirm it and keep the rest of your record right.
Claim this profileAcademic lineage
View as a treeAdvisors
Works62 from public data
- Visual Commonsense R-CNN299
A novel unsupervised feature representation learning method, Visual Commonsense Region-based Convolutional Neural Network (VC R-CNN), is presented to serve as an improved visual region encoder for high-level tasks such as captioning and VQA, and observes consistent performance boosts across them.
- 282
- Counterfactual Zero-Shot and Open-Set Visual Recognition249
A novel counterfactual framework for both Zero-Shot Learning (ZSL) and Open-Set Recognition (OSR), whose common challenge is generalizing to the unseen-classes by only training on the seen-classes is presented.
- Class Re-Activation Maps for Weakly-Supervised Semantic Segmentation235
An embarrassingly simple yet surprisingly effective method: Reactivating the converged CAM with BCE by using softmax crossentropy loss (SCE), dubbed ReCAM, which not only generates high-quality masks, but also supports plug-and-play in any CAM variant with little overhead.
- 197
- Disco: Disentangled Control for Realistic Human Dance Generation187
Disco is introduced, which includes a novel model architecture with disentangled control to improve the compositionality of dance synthesis, and an effective human attribute pre-training for better generalizability to unseen humans.
- Matching Images and Text with Multi-modal Tensor Fusion and Re-ranking177
This work proposes a novel Multi-modal Tensor Fusion Network (MTFN) to explicitly learn an accurate image-text similarity function with rank-based tensor fusion rather than seeking a common embedding space for each image- text instance.
- Causal Attention for Unbiased Visual Recognition174
A causal attention module (CaaM) that self-annotates the confounders in unsupervised fashion is proposed that can be stacked and integrated in conventional attention CNN and self-attention Vision Transformer, and in OOD settings, deep models with CaaM outperform those without it significantly.
- Conditional preferences and updating112
This paper axiomatizes updating rules for preferences that are not necessarily in the expected utility class: the traditional BayesRule, the Dempster–Shafer rule, and the generalized Bayes rule.
- 111
- 94
- A Novel Pure Pursuit Algorithm for Autonomous Vehicles Based on Salp Swarm Algorithm and Velocity Controller88
A novel pure pursuit algorithm based on the optimized look-ahead distance named OLDPPA is proposed and experimental results show that the tracking accuracy of OL DPPA is better than other algorithms.
- An improved African vultures optimization algorithm based on tent chaotic mapping and time-varying mechanism81
A new metaheuristic algorithm, an improved African vultures optimization algorithm based on tent chaotic mapping and time-varying mechanism (TAVOA), is proposed, which shows good performance in real-world engineering design problems.
- Equivariant Similarity for Vision-Language Foundation Models74
This study proposes EqSim, a regularization loss that can be efficiently calculated from any two matched training pairs and easily pluggable into existing image-text retrieval fine-tuning and presents a new challenging benchmark EqBen, the first to focus on "visual-minimal change".
- A novel filter feature selection algorithm based on relief74
A novel feature selection algorithm, named Multidirectional Relief (MRelief), is proposed, which is significantly better than other algorithms including LPLIR, ReliefF, LLH-Relief, MultiSURF, MSLIR-NN, MRMR, MPMD and STIR in this study.
- 74
- Impact of Social Participation Types on Depression in the Elderly in China: An Analysis Based on Counterfactual Causal Inference71
Through PSM analysis, different types of social participation have different effects on depression in the elderly, and it is necessary to provide differentiated social support environment for the elderly to maximize the positive effects of different typesof social participation.
- Enhance Image Classification via Inter-Class Image Mixup with Diffusion Model60
The empirical results demonstrate that Diff-Mix achieves a better balance between faith-fulness and diversity, leading to a marked improvement in performance across diverse image classification scenarios, including few-shot, conventional, and long-tail classifications for domain-specific datasets.
- ClothFormer: Taming Video Virtual Try-on in All Module45
A novel video virtual try-on framework, ClothFormer, which successfully synthesizes realistic, harmonious, and spatio-temporal consistent results in complicated environment is proposed.
- 43
- A Hybrid Improved Dragonfly Algorithm for Feature Selection41
A novel feature selection algorithm named hybrid improved dragonflies algorithm (HIDA) which combines the advantages of both mRMR and improved dragonfly algorithm (IDA) in order to generate promising candidate subset and achieve higher classification accuracy rate.
- 34
- Equivariance and Invariance Inductive Bias for Learning from Insufficient Data26
State-of-the-art experimental results on real-world benchmarks (VIPriors, ImageNet100 and NICO) validate the great potential of equivariance and invariance in data-efficient learning.
- 26
- Pricing of New Securities in an Incomplete Market: the Catch 22 of No‐Arbitrage Pricing18
This paper provides a simple proof that the introduction of the new security will typically change the prices of all the existing securities and clarifies the distinction between the two approaches to the valuation of a new security in an incomplete market.
- A Novel Network Architecture of Decision-Making for Self-Driving Vehicles Based on Long Short-Term Memory and Grasshopper Optimization Algorithm17
Improved long short-term memory network with support vector machine classifier optimized by grasshopper optimization algorithm (GOA-ImLSTM) improves the accuracy of results compared with other decision-making methods for self-driving vehicles on the Next Generation SIMulation.
- 16
- A Novel Gated Recurrent Unit Network Based on SVM and Moth-Flame Optimization Algorithm for Behavior Decision-Making of Autonomous Vehicles14
The experimental results on the NGSIM dataset show that EnMFO-ImGRU brings higher accuracy than existing methods for the behavior decision-making results of autonomous vehicles.
- 12
- Global chaotic bat algorithm for feature selection11
A global chaotic bat algorithm (GCBA), which applies chaotic map to population initialization to cover the entire solution space and an improved transfer function is proposed, which transfers the continuous space to discrete binary space.
- 10
- Explaining Language Models' Predictions with High-Impact Concepts9
This work proposes a post-hoc interpretability method for extracting predictive high-level features (concepts) from the pretrained model's hidden layer activations and optimize for features whose existence causes the output predictions to change substantially, which generates a high impact.
- 9
- 7
- 6
- 5
- 5
- 4
- 4
- On condition numbers in the cyclic reduction processes of a tridiagonal matrix4
It is proved that the condition number for eigenvalues of some classes of coefficient matrices always decreases after the first step of the cyclic reduction method.
- 3
- Spectrum Sharing for 5G3
It is concluded that spectrum sharing becomes more and more important for 5G systems as a complementary way of using spectrum.
- 3
- 2
- 1
- 1
- DyFract: Dynamic fractal embedding with semantic constraints for temporal knowledge graph completion–
- –
- Incorporating Hybrid Prediction with Feature Selection to Estimate Carbon Emissions with Limited Data–
The Likelihood Ratio (LR) test, the soundness test, and the heterogeneity test have confirmed the generalizability and stability of the proposed hybrid model for CO 2 emissions predictions, as well as the effectiveness of feature selection.
- –
- –
- –
- –
- –
- –
- –
- –
- –
- Communication Using a Reconfigurable and Reliable Transport Layer Protocol–
The outcome of the research is a reconfigurable, user-level, reliable transport layer protocol, called RRTP (Reliable and Reconfigurable Transport Protocol) that is TCP-friendly, i.e. it asymptotically converges to fairness as in the case of LIMD (Linear Increase Multiplicative Decrease) algorithms.
- A User Level, Reliable, and Reconfigurable Transport Layer Protocol–
This paper explores an alternative transport layer protocol that is more suitable for today’s mobile as well as other non-conventional network environments, called RRTP (Reliable and Reconfigurable Transport Protocol), which runs atop of UDP.
- A MEASUREMENT SYSTEM OF HUMAN-BODY CONSTITUTION AND STRENGTH EXERCISE–
This system adopts the sensor technology, computer software and hardware technology, as well as digital display technology to perform automatic data collection and processing in real time in order to enhance the system accuracy and the running efficiency.
- –
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 2026-10-12. 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.