Publications (63)
Token Boosting for Robust Self-Supervised Visual Transformer Pre-training
Tianjiao Li, Lin Geng Foo, Ping Hu +4
Learning with large-scale unlabeled data has become a powerful tool for pre-training Visual Transformers (VTs). However, prior works tend to overlook that, in real-world scenarios,…
Deep Generative Modeling for Financial Time Series with Application in VaR: A Comparative Review
Lars Ericson, Xuejun Zhu, Xusi Han +4
In the financial services industry, forecasting the risk factor distribution conditional on the history and the current market environment is the key to market risk modeling in gen…
Splatting-based Synthesis for Video Frame Interpolation
Simon Niklaus, Ping Hu, Jiawen Chen
Frame interpolation is an essential video processing technique that adjusts the temporal resolution of an image sequence. While deep learning has brought great improvements to the…
Tilings in graphons
Jan Hladky, Ping Hu, Diana Piguet
We introduce a counterpart to the notion of vertex disjoint tilings by copy of a fixed graph F to the setting of graphons. The case F=K_2 gives the notion of matchings in graphons.…
On the Turán number of double stars
Ping Hu, Ting Lan
The Turán number of a graph , , is the maximum number of edges in a graph on vertices which does not contain as a subgraph. Let denote a double star…
Non-Asymptotic Performance of Social Machine Learning Under Limited Data
Ping Hu, Virginia Bordignon, Mert Kayaalp +1
This paper studies the probability of error associated with the social machine learning framework, which involves an independent training phase followed by a cooperative decision-m…
Diffusion-based Image Translation with Label Guidance for Domain Adaptive Semantic Segmentation
Duo Peng, Ping Hu, Qiuhong Ke +1
Translating images from a source domain to a target domain for learning target models is one of the most common strategies in domain adaptive semantic segmentation (DASS). However,…
Minimum number of edges that occur in odd cycles
Andrzej Grzesik, Ping Hu, Jan Volec
If a graph has vertices and more than edges, then it contains a copy of . In 1992, ErdÅs, Faudree and Rousseau showed even more, that the number of edge…
2023 Low-Power Computer Vision Challenge (LPCVC) Summary
Leo Chen, Benjamin Boardley, Ping Hu +27
This article describes the 2023 IEEE Low-Power Computer Vision Challenge (LPCVC). Since 2015, LPCVC has been an international competition devoted to tackling the challenge of compu…
Graph Smoothing for Enhanced Local Geometry Learning in Point Cloud Analysis
Shangbo Yuan, Jie Xu, Ping Hu +2
Graph-based methods have proven to be effective in capturing relationships among points for 3D point cloud analysis. However, these methods often suffer from suboptimal graph struc…
RADAR: A Risk-Aware Dynamic Multi-Agent Framework for LLM Safety Evaluation via Role-Specialized Collaboration
Xiuyuan Chen, Jian Zhao, Yuchen Yuan +8
Existing safety evaluation methods for large language models (LLMs) suffer from inherent limitations, including evaluator bias and detection failures arising from model homogeneity…
ZeroWaste Dataset: Towards Deformable Object Segmentation in Cluttered Scenes
Dina Bashkirova, Mohamed Abdelfattah, Ziliang Zhu +7
Less than 35% of recyclable waste is being actually recycled in the US, which leads to increased soil and sea pollution and is one of the major concerns of environmental researcher…
Weakly-supervised Compositional FeatureAggregation for Few-shot Recognition
Ping Hu, Ximeng Sun, Kate Saenko +1
Learning from a few examples is a challenging task for machine learning. While recent progress has been made for this problem, most of the existing methods ignore the compositional…
On the Chromatic Thresholds of Hypergraphs
József Balogh, Jane Butterfield, Ping Hu +2
Let F be a family of r-uniform hypergraphs. The chromatic threshold of F is the infimum of all non-negative reals c such that the subfamily of F comprising hypergraphs H with minim…
Maximum density of an induced 5-cycle is achieved by an iterated blow-up of a 5-cycle
József Balogh, Ping Hu, Bernard Lidický +1
Let denote the maximum number of induced copies of 5-cycles in graphs on vertices. For large enough, we show that $C(n)=a\cdot b\cdot c \cdot d \cdot e + C(a)+C(b)+C…
On Efficient Variants of Segment Anything Model: A Survey
Xiaorui Sun, Jun Liu, Heng Tao Shen +2
The Segment Anything Model (SAM) is a foundational model for image segmentation tasks, known for its strong generalization across diverse applications. However, its impressive perf…
Skew-symmetric Nitsche's formulation in isogeometric analysis: Dirichlet and symmetry conditions, patch coupling and frictionless contact
Qingyuan Hu, Franz Chouly, Ping Hu +2
A simple skew-symmetric Nitsche's formulation is introduced into the framework of isogeometric analysis (IGA) to deal with various problems in small strain elasticity: essential bo…
Adaptive Multi-Modality Prompt Learning
Zongqian Wu, Yujing Liu, Mengmeng Zhan +3
Although current prompt learning methods have successfully been designed to effectively reuse the large pre-trained models without fine-tuning their large number of parameters, the…
The inducibility of oriented stars
Ping Hu, Jie Ma, Sergey Norin +1
We consider the problem of maximizing the number of induced copies of an oriented star in digraphs of given size, where the center of the star has out-degree and i…
Komlós's tiling theorem via graphon covers
Jan Hladký, Ping Hu, Diana Piguet
Komlos [Komlos: Tiling Turan Theorems, Combinatorica, 2000] determined the asymptotically optimal minimum-degree condition for covering a given proportion of vertices of a host gra…
Isogeometric analysis of thin Reissner-Mindlin plates and shells: locking phenomena and B-bar method
Qingyuan Hu, Yang Xia, Sundararajan Natarajan +3
We propose a local type of B-bar formulation, addressing locking in degenerated Reissner-Mindlin plate and shell formulations in the context of isogeometric analysis. Parasitic str…
DisasterBench: A Multimodal Benchmark for UAV-Based Disaster Response in Complex Environments
Tan Zhang, Quanyou Li, Lu Zhang +3
When a disaster unfolds, responders must answer not only what is happening, but also why it is happening, what will happen next, and what to do now, often from noisy low-altitude U…
Deepsea: A Meta-ocean Prototype for Undersea Exploration
Jinyu Li, Ping Hu, Weicheng Cui +2
Metaverse has attracted great attention from industry and academia in recent years. Metaverse for the ocean (Meta-ocean) is the implementation of the Metaverse technologies in virt…
Rainbow triangles in three-colored graphs
Jozsef Balogh, Ping Hu, Bernard Lidicky +3
Erdos and Sos proposed a problem of determining the maximum number F(n) of rainbow triangles in 3-edge-colored complete graphs on n vertices. They conjectured that F(n) = F(a)+ F(b…
DualCoOp: Fast Adaptation to Multi-Label Recognition with Limited Annotations
Ximeng Sun, Ping Hu, Kate Saenko
Solving multi-label recognition (MLR) for images in the low-label regime is a challenging task with many real-world applications. Recent work learns an alignment between textual an…
Automatic Method Illustration Generation for AI Scientific Papers via Drawing Middleware Creation, Evolution, and Orchestration
Zhuoling Li, Jiarui Zhang, Ping Hu +4
Method illustrations (MIs) play a crucial role in conveying the core ideas of scientific papers, yet their generation remains a labor-intensive process. Here, we take inspiration f…
Noisy Node Classification by Bi-level Optimization based Multi-teacher Distillation
Yujing Liu, Zongqian Wu, Zhengyu Lu +4
Previous graph neural networks (GNNs) usually assume that the graph data is with clean labels for representation learning, but it is not true in real applications. In this paper, w…
Boosting Continual Learning of Vision-Language Models via Mixture-of-Experts Adapters
Jiazuo Yu, Yunzhi Zhuge, Lu Zhang +4
Continual learning can empower vision-language models to continuously acquire new knowledge, without the need for access to the entire historical dataset. However, mitigating the p…
Submerse: Visualizing Storm Surge Flooding Simulations in Immersive Display Ecologies
Saeed Boorboor, Yoonsang Kim, Ping Hu +3
We present Submerse, an end-to-end framework for visualizing flooding scenarios on large and immersive display ecologies. Specifically, we reconstruct a surface mesh from input flo…
Real-time Semantic Segmentation with Fast Attention
Ping Hu, Federico Perazzi, Fabian Caba Heilbron +4
In deep CNN based models for semantic segmentation, high accuracy relies on rich spatial context (large receptive fields) and fine spatial details (high resolution), both of which…
FineRS: Fine-grained Reasoning and Segmentation of Small Objects with Reinforcement Learning
Lu Zhang, Jiazuo Yu, Haomiao Xiong +4
Multi-modal Large Language Models (MLLMs) have shown remarkable capabilities across a wide range of vision-language tasks. However, due to the restricted input resolutions, MLLMs f…
Unified Prompt Attack Against Text-to-Image Generation Models
Duo Peng, Qiuhong Ke, Mark He Huang +2
Text-to-Image (T2I) models have advanced significantly, but their growing popularity raises security concerns due to their potential to generate harmful images. To address these is…
Upper bounds on the size of 4- and 6-cycle-free subgraphs of the hypercube
József Balogh, Ping Hu, Bernard Lidický +1
In this paper we modify slightly Razborov's flag algebra machinery to be suitable for the hypercube. We use this modified method to show that the maximum number of edges of a 4-cyc…
Cross-Domain Few-Shot Learning via Multi-View Collaborative Optimization with Vision-Language Models
Dexia Chen, Wentao Zhang, Qianjie Zhu +4
Vision-language models (VLMs) pre-trained on natural image and language data, such as CLIP, have exhibited significant potential in few-shot image recognition tasks, leading to dev…
Towards Generalized Range-View LiDAR Segmentation in Adverse Weather
Longyu Yang, Lu Zhang, Jun Liu +4
LiDAR segmentation has emerged as an important task to enrich scene perception and understanding. Range-view-based methods have gained popularity due to their high computational ef…
Enhancing Few-Shot Out-of-Distribution Detection via the Refinement of Foreground and Background
Tianyu Li, Zongqian Wu, Songyue Cai +2
CLIP-based foreground-background (FG-BG) decomposition methods have demonstrated remarkable effectiveness in improving few-shot out-of-distribution (OOD) detection performance. How…
Background Prompt for Few-Shot Out-of-Distribution Detection
Songyue Cai, Zongqian Wu, Yujie Mo +4
Existing foreground-background (FG-BG) decomposition methods for the few-shot out-of-distribution (FS-OOD) detection often suffer from low robustness due to over-reliance on the lo…
TSTMotion: Training-free Scene-aware Text-to-motion Generation
Ziyan Guo, Haoxuan Qu, Hossein Rahmani +4
Text-to-motion generation has recently garnered significant research interest, primarily focusing on generating human motion sequences in blank backgrounds. However, human motions…
Phase transitions in the Ramsey-Turán theory
József Balogh, Ping Hu, Miklós Simonovits
Let be a function and be a graph. Denote by the maximum number of edges of an -free graph on vertices with independence number less than . Er…
Broadcast Repair for Wireless Distributed Storage Systems
Ping Hu, Chi Wan Sung, Terence H. Chan
In wireless distributed storage systems, storage nodes are connected by wireless channels, which are broadcast in nature. This paper exploits this unique feature to design an effic…
LLM-based Agentic Reasoning Frameworks: A Survey from Methods to Scenarios
Bingxi Zhao, Lin Geng Foo, Ping Hu +3
Recent advances in the intrinsic reasoning capabilities of large language models (LLMs) have given rise to LLM-based agent systems that exhibit near-human performance on a variety…
Conterfactual Generative Zero-Shot Semantic Segmentation
Feihong Shen, Jun Liu, Ping Hu
zero-shot learning is an essential part of computer vision. As a classical downstream task, zero-shot semantic segmentation has been studied because of its applicant value. One of…
DualCoOp++: Fast and Effective Adaptation to Multi-Label Recognition with Limited Annotations
Ping Hu, Ximeng Sun, Stan Sclaroff +1
Multi-label image recognition in the low-label regime is a task of great challenge and practical significance. Previous works have focused on learning the alignment between textual…
Optimal Aggregation Strategies for Social Learning over Graphs
Ping Hu, Virginia Bordignon, Stefan Vlaski +1
Adaptive social learning is a useful tool for studying distributed decision-making problems over graphs. This paper investigates the effect of combination policies on the performan…
AsynFusion: Towards Asynchronous Latent Consistency Models for Decoupled Whole-Body Audio-Driven Avatars
Tianbao Zhang, Jian Zhao, Yuer Li +5
Whole-body audio-driven avatar pose and expression generation is a critical task for creating lifelike digital humans and enhancing the capabilities of interactive virtual agents,…
Koala: Key frame-conditioned long video-LLM
Reuben Tan, Ximeng Sun, Ping Hu +5
Long video question answering is a challenging task that involves recognizing short-term activities and reasoning about their fine-grained relationships. State-of-the-art video Lar…
Minimum number of monotone subsequences of length 4 in permutations
József Balogh, Ping Hu, Bernard Lidický +3
We show that for every sufficiently large , the number of monotone subsequences of length four in a permutation on points is at least $\binom{\lfloor n/3 \rfloor}{4} + \bino…
DreamMix: Decoupling Object Attributes for Enhanced Editability in Customized Image Inpainting
Yicheng Yang, Pengxiang Li, Lu Zhang +6
Subject-driven image inpainting has recently gained prominence in image editing with the rapid advancement of diffusion models. Beyond image guidance, recent studies have explored…
Learning to Detect Every Thing in an Open World
Kuniaki Saito, Ping Hu, Trevor Darrell +1
Many open-world applications require the detection of novel objects, yet state-of-the-art object detection and instance segmentation networks do not excel at this task. The key iss…
Capacity of Wireless Distributed Storage Systems with Broadcast Repair
Ping Hu, Chi Wan Sung, Terence H. Chan
In wireless distributed storage systems, storage nodes are connected by wireless channels, which are broadcast in nature. This paper exploits this unique feature to design an effic…
Modality-Aware Feature Matching in Visual and Vision-Language Applications: A Comprehensive Survey
Weide Liu, Wei Zhou, Jun Liu +4
Feature matching is a cornerstone task in computer vision, essential for applications such as image retrieval, stereo matching, 3D reconstruction, and SLAM. This survey comprehensi…
The List Linear Arboricity of Digraphs
Yueping Shi, Ping Hu
A (directed) linear forest is a (di)graph whose components are (directed) paths. The linear arboricity of a (di)graph is the minimum number of (directed) linear forests…
Mantel's Theorem for Random Hypergraphs
József Balogh, Jane Butterfield, Ping Hu +1
A classical result in extremal graph theory is Mantel's Theorem, which states that every maximum triangle-free subgraph of is bipartite. A sparse version of Mantel's Theorem…
Towards Explicit Geometry-Reflectance Collaboration for Generalized LiDAR Segmentation in Adverse Weather
Longyu Yang, Ping Hu, Shangbo Yuan +4
Existing LiDAR semantic segmentation models often suffer from decreased accuracy when exposed to adverse weather conditions. Recent methods addressing this issue focus on enhancing…
Video Frame Interpolation with Many-to-many Splatting and Spatial Selective Refinement
Ping Hu, Simon Niklaus, Lu Zhang +2
In this work, we first propose a fully differentiable Many-to-Many (M2M) splatting framework to interpolate frames efficiently. Given a frame pair, we estimate multiple bidirection…
E-VAds: An E-commerce Short Videos Understanding Benchmark for MLLMs
Xianjie Liu, Yiman Hu, Liang Wu +4
E-commerce short videos represent a high-revenue segment of the online video industry characterized by a goal-driven format and dense multi-modal signals. Current models often stru…
Geometry-Aware Planar Embedding of Treelike Structures
Ping Hu, Saeed Boorboor, Joseph Marino +1
The growing complexity of spatial and structural information in 3D data makes data inspection and visualization a challenging task. We describe a method to create a planar embeddin…
Bootstraping Clustering of Gaussians for View-consistent 3D Scene Understanding
Wenbo Zhang, Lu Zhang, Ping Hu +3
Injecting semantics into 3D Gaussian Splatting (3DGS) has recently garnered significant attention. While current approaches typically distill 3D semantic features from 2D foundatio…
Many-to-many Splatting for Efficient Video Frame Interpolation
Ping Hu, Simon Niklaus, Stan Sclaroff +1
Motion-based video frame interpolation commonly relies on optical flow to warp pixels from the inputs to the desired interpolation instant. Yet due to the inherent challenges of mo…
Temporally Distributed Networks for Fast Video Semantic Segmentation
Ping Hu, Fabian Caba Heilbron, Oliver Wang +3
We present TDNet, a temporally distributed network designed for fast and accurate video semantic segmentation. We observe that features extracted from a certain high-level layer of…
Clique-factors in graphs with sublinear -independence number
Jie Han, Ping Hu, Guanghui Wang +1
Given a graph and an integer , we denote by the maximum size of a -free subset of vertices in . A recent question of Nenadov and Pehov…
The Final Layer Holds the Key: A Unified and Efficient GNN Calibration Framework
Jincheng Huang, Jie Xu, Xiaoshuang Shi +3
Graph Neural Networks (GNNs) have demonstrated remarkable effectiveness on graph-based tasks. However, their predictive confidence is often miscalibrated, typically exhibiting unde…
Densities of 3-vertex graphs
Roman Glebov, Andrzej Grzesik, Ping Hu +3
Let d_i(G) be the density of the 3-vertex i-edge graph in a graph G, i.e., the probability that three random vertices induce a subgraph with i edges. Let S be the set of all quadru…