Search Results for author: Tianyu Wang

Found 57 papers, 17 papers with code

Spatial Attentive Single-Image Deraining with a High Quality Real Rain Dataset

2 code implementations CVPR 2019 Tianyu Wang, Xin Yang, Ke Xu, Shaozhe Chen, Qiang Zhang, Rynson Lau

Second, to better cover the stochastic distribution of real rain streaks, we propose a novel SPatial Attentive Network (SPANet) to remove rain streaks in a local-to-global manner.

Single Image Deraining Vocal Bursts Intensity Prediction

Instance Shadow Detection

3 code implementations CVPR 2020 Tianyu Wang, Xiao-Wei Hu, Qiong Wang, Pheng-Ann Heng, Chi-Wing Fu

Then, we pair up the predicted shadow and object instances, and match them with the predicted shadow-object associations to generate the final results.

Instance Shadow Detection Object +1

Instance Shadow Detection with A Single-Stage Detector

2 code implementations11 Jul 2022 Tianyu Wang, Xiaowei Hu, Pheng-Ann Heng, Chi-Wing Fu

This paper formulates a new problem, instance shadow detection, which aims to detect shadow instance and the associated object instance that cast each shadow in the input image.

Instance Shadow Detection Object +2

Sparse2Dense: Learning to Densify 3D Features for 3D Object Detection

1 code implementation23 Nov 2022 Tianyu Wang, Xiaowei Hu, Zhengzhe Liu, Chi-Wing Fu

Importantly, we formulate the lightweight plug-in S2D module and the point cloud reconstruction module in SDet to densify 3D features and train SDet to produce 3D features, following the dense 3D features in DDet.

3D Object Detection Domain Adaptation +2

Image sensing with multilayer, nonlinear optical neural networks

1 code implementation27 Jul 2022 Tianyu Wang, Mandar M. Sohoni, Logan G. Wright, Martin M. Stein, Shi-Yuan Ma, Tatsuhiro Onodera, Maxwell G. Anderson, Peter L. McMahon

In image sensing, a measurement, such as of an object's position, is performed by computational analysis of a digitized image.

Image Classification

Revisiting Shadow Detection: A New Benchmark Dataset for Complex World

2 code implementations16 Nov 2019 Xiaowei Hu, Tianyu Wang, Chi-Wing Fu, Yitong Jiang, Qiong Wang, Pheng-Ann Heng

Shadow detection in general photos is a nontrivial problem, due to the complexity of the real world.

Shadow Detection

An optical neural network using less than 1 photon per multiplication

2 code implementations27 Apr 2021 Tianyu Wang, Shi-Yuan Ma, Logan G. Wright, Tatsuhiro Onodera, Brian Richard, Peter L. McMahon

Here, we experimentally demonstrate an optical neural network achieving 99% accuracy on handwritten-digit classification using ~3. 2 detected photons per weight multiplication and ~90% accuracy using ~0. 64 photons (~$2. 4 \times 10^{-19}$ J of optical energy) per weight multiplication.

Total Energy

SILT: Shadow-aware Iterative Label Tuning for Learning to Detect Shadows from Noisy Labels

1 code implementation ICCV 2023 Han Yang, Tianyu Wang, Xiaowei Hu, Chi-Wing Fu

Existing shadow detection datasets often contain missing or mislabeled shadows, which can hinder the performance of deep learning models trained directly on such data.

Shadow Detection

WYWEB: A NLP Evaluation Benchmark For Classical Chinese

1 code implementation23 May 2023 Bo Zhou, Qianglong Chen, Tianyu Wang, Xiaomi Zhong, Yin Zhang

To fully evaluate the overall performance of different NLP models in a given domain, many evaluation benchmarks are proposed, such as GLUE, SuperGLUE and CLUE.

Machine Translation Natural Language Understanding +2

Adaptive Discretization for Model-Based Reinforcement Learning

1 code implementation NeurIPS 2020 Sean R. Sinclair, Tianyu Wang, Gauri Jain, Siddhartha Banerjee, Christina Lee Yu

We introduce the technique of adaptive discretization to design an efficient model-based episodic reinforcement learning algorithm in large (potentially continuous) state-action spaces.

Model-based Reinforcement Learning reinforcement-learning +1

Conceptual similarity and communicative need shape colexification: an experimental study

1 code implementation19 Mar 2021 Andres Karjus, Richard A. Blythe, Simon Kirby, Tianyu Wang, Kenny Smith

Colexification refers to the phenomenon of multiple meanings sharing one word in a language.

Scaling on-chip photonic neural processors using arbitrarily programmable wave propagation

1 code implementation27 Feb 2024 Tatsuhiro Onodera, Martin M. Stein, Benjamin A. Ash, Mandar M. Sohoni, Melissa Bosch, Ryotatsu Yanagimoto, Marc Jankowski, Timothy P. McKenna, Tianyu Wang, Gennady Shvets, Maxim R. Shcherbakov, Logan G. Wright, Peter L. McMahon

On-chip photonic processors for neural networks have potential benefits in both speed and energy efficiency but have not yet reached the scale at which they can outperform electronic processors.

Vowel Classification

On Sharp Stochastic Zeroth Order Hessian Estimators over Riemannian Manifolds

1 code implementation26 Jan 2022 Tianyu Wang

We show that, for an analytic real-valued function $f$, our estimator achieves a bias bound of order $ O \left( \gamma \delta^2 \right) $, where $ \gamma $ depends on both the Levi-Civita connection and function $f$, and $\delta$ is the finite difference step size.

Towards Practical Lipschitz Bandits

no code implementations26 Jan 2019 Tianyu Wang, Weicheng Ye, Dawei Geng, Cynthia Rudin

Stochastic Lipschitz bandit algorithms balance exploration and exploitation, and have been used for a variety of important task domains.

Gaussian Processes

SAC-Net: Spatial Attenuation Context for Salient Object Detection

no code implementations25 Mar 2019 Xiaowei Hu, Chi-Wing Fu, Lei Zhu, Tianyu Wang, Pheng-Ann Heng

This paper presents a new deep neural network design for salient object detection by maximizing the integration of local and global image context within, around, and beyond the salient objects.

Object object-detection +2

Large Scale Model Predictive Control with Neural Networks and Primal Active Sets

no code implementations23 Oct 2019 Steven W. Chen, Tianyu Wang, Nikolay Atanasov, Vijay Kumar, Manfred Morari

The approach combines an offline-trained fully-connected neural network with an online primal active set solver.

Model Predictive Control

Transfer Learning for Hate Speech Detection in Social Media

no code implementations10 Jun 2019 Marian-Andrei Rizoiu, Tianyu Wang, Gabriela Ferraro, Hanna Suominen

This paper uses a transfer learning technique to leverage two independent datasets jointly and builds a single representation of hate speech.

Social and Information Networks Computers and Society

Learning Navigation Costs from Demonstration in Partially Observable Environments

no code implementations26 Feb 2020 Tianyu Wang, Vikas Dhiman, Nikolay Atanasov

This paper focuses on inverse reinforcement learning (IRL) to enable safe and efficient autonomous navigation in unknown partially observable environments.

Autonomous Navigation Motion Planning +1

Learning Navigation Costs from Demonstration with Semantic Observations

no code implementations9 Jun 2020 Tianyu Wang, Vikas Dhiman, Nikolay Atanasov

The objective is to infer a cost function that explains demonstrated behavior while relying only on the expert's observations and state-control trajectory.

Autonomous Driving Motion Planning +1

Bandits for BMO Functions

no code implementations ICML 2020 Tianyu Wang, Cynthia Rudin

We study the bandit problem where the underlying expected reward is a Bounded Mean Oscillation (BMO) function.

Episodic Linear Quadratic Regulators with Low-rank Transitions

no code implementations3 Nov 2020 Tianyu Wang, Lin F. Yang

Consequently, the sample complexity of our algorithm only depends on the rank, $m$, rather than the ambient dimension, $d$, which can be orders-of-magnitude larger.

Towards Fundamental Limits of Multi-armed Bandits with Random Walk Feedback

no code implementations3 Nov 2020 Tianyu Wang, Lin F. Yang, Zizhuo Wang

In this paper, we consider a new Multi-Armed Bandit (MAB) problem where arms are nodes in an unknown and possibly changing graph, and the agent (i) initiates random walks over the graph by pulling arms, (ii) observes the random walk trajectories, and (iii) receives rewards equal to the lengths of the walks.

Multi-Armed Bandits Recommendation Systems

Inverse reinforcement learning for autonomous navigation via differentiable semantic mapping and planning

no code implementations1 Jan 2021 Tianyu Wang, Vikas Dhiman, Nikolay Atanasov

The objective is to infer a cost function that explains demonstrated behavior while relying only on the expert's observations and state-control trajectory.

Autonomous Driving Autonomous Navigation +3

Spatially Invariant Unsupervised 3D Object-Centric Learning and Scene Decomposition

no code implementations10 Jun 2021 Tianyu Wang, Miaomiao Liu, Kee Siong Ng

Experimental results demonstrate that SPAIR3D has strong scalability and is capable of detecting and segmenting an unknown number of objects from a point cloud in an unsupervised manner.

Object Relational Reasoning +1

Distributionally Robust Prescriptive Analytics with Wasserstein Distance

no code implementations10 Jun 2021 Tianyu Wang, Ningyuan Chen, Chun Wang

In prescriptive analytics, the decision-maker observes historical samples of $(X, Y)$, where $Y$ is the uncertain problem parameter and $X$ is the concurrent covariate, without knowing the joint distribution.

Portfolio Optimization

From the Greene--Wu Convolution to Gradient Estimation over Riemannian Manifolds

no code implementations17 Aug 2021 Tianyu Wang, Yifeng Huang, Didong Li

Over a complete Riemannian manifold of finite dimension, Greene and Wu introduced a convolution, known as Greene-Wu (GW) convolution.

Towards Accurate Alignment in Real-time 3D Hand-Mesh Reconstruction

no code implementations ICCV 2021 Xiao Tang, Tianyu Wang, Chi-Wing Fu

3D hand-mesh reconstruction from RGB images facilitates many applications, including augmented reality (AR).

Lipschitz Bandits with Batched Feedback

no code implementations19 Oct 2021 Yasong Feng, Zengfeng Huang, Tianyu Wang

Specifically, we show that for a $T$-step problem with Lipschitz reward of zooming dimension $d_z$, our algorithm achieves theoretically optimal (up to logarithmic factors) regret rate $\widetilde{\mathcal{O}}\left(T^{\frac{d_z+1}{d_z+2}}\right)$ using only $ \mathcal{O} \left( \log\log T\right) $ batches.

Learning Navigation Costs from Demonstrations with Semantic Observations

no code implementations L4DC 2020 Tianyu Wang, Vikas Dhiman, Nikolay Atanasov

The objective is to infer a cost function that explains demonstrated behavior while relying only on the expert’s observations and state-control trajectory.

Autonomous Driving Motion Planning +1

DMF-Net: Dual-Branch Multi-Scale Feature Fusion Network for copy forgery identification of anti-counterfeiting QR code

no code implementations19 Jan 2022 Zhongyuan Guo, Hong Zheng, Changhui You, Tianyu Wang, Chang Liu

We first analyze the production principle of anti-counterfeiting QR code, and convert the identification of copy forgery to device category forensics, and then a Dual-Branch Multi-Scale Feature Fusion network is proposed.

Image Forensics

Stochastic Zeroth Order Gradient and Hessian Estimators: Variance Reduction and Refined Bias Bounds

1 code implementation29 May 2022 Yasong Feng, Tianyu Wang

In particular, we design estimators for smooth functions such that, if one uses $ \Theta \left( k \right) $ random directions sampled from the Stiefel's manifold $ \text{St} (n, k) $ and finite-difference granularity $\delta$, the variance of the gradient estimator is bounded by $ \mathcal{O} \left( \left( \frac{n}{k} - 1 \right) + \left( \frac{n^2}{k} - n \right) \delta^2 + \frac{ n^2 \delta^4 }{ k } \right) $, and the variance of the Hessian estimator is bounded by $\mathcal{O} \left( \left( \frac{n^2}{k^2} - 1 \right) + \left( \frac{n^4}{k^2} - n^2 \right) \delta^2 + \frac{n^4 \delta^4 }{k^2} \right) $.

Latent Policies for Adversarial Imitation Learning

no code implementations22 Jun 2022 Tianyu Wang, Nikhil Karnwal, Nikolay Atanasov

We use an action encoder-decoder model to obtain a low-dimensional latent action space and train a LAtent Policy using Adversarial imitation Learning (LAPAL).

Imitation Learning

A Direct Approximation of AIXI Using Logical State Abstractions

no code implementations13 Oct 2022 Samuel Yang-Zhao, Tianyu Wang, Kee Siong Ng

We propose a practical integration of logical state abstraction with AIXI, a Bayesian optimality notion for reinforcement learning agents, to significantly expand the model class that AIXI agents can be approximated over to complex history-dependent and structured environments.

Convergence Rates of Stochastic Zeroth-order Gradient Descent for Ł ojasiewicz Functions

no code implementations31 Oct 2022 Tianyu Wang, Yasong Feng

We prove convergence rates of Stochastic Zeroth-order Gradient Descent (SZGD) algorithms for Lojasiewicz functions.

Linear RNNs Provably Learn Linear Dynamic Systems

no code implementations19 Nov 2022 Lifu Wang, Tianyu Wang, Shengwei Yi, Bo Shen, Bo Hu, Xing Cao

We study the learning ability of linear recurrent neural networks with Gradient Descent.

Video Instance Shadow Detection

no code implementations23 Nov 2022 Zhenghao Xing, Tianyu Wang, Xiaowei Hu, Haoran Wu, Chi-Wing Fu, Pheng-Ann Heng

First, we design SSIS-Track, a new framework to extract shadow-object associations in videos with paired tracking and without category specification; especially, we strive to maintain paired tracking even the objects/shadows are temporarily occluded for several frames.

Instance Shadow Detection Shadow Detection

Hedging Complexity in Generalization via a Parametric Distributionally Robust Optimization Framework

no code implementations3 Dec 2022 Garud Iyengar, Henry Lam, Tianyu Wang

We propose a simple approach in which the distribution of random perturbations is approximated using a parametric family of distributions.

Generalization Bounds Management +2

Optical Transformers

no code implementations20 Feb 2023 Maxwell G. Anderson, Shi-Yuan Ma, Tianyu Wang, Logan G. Wright, Peter L. McMahon

We conclude that with well-engineered, large-scale optical hardware, it may be possible to achieve a $100 \times$ energy-efficiency advantage for running some of the largest current Transformer models, and that if both the models and the optical hardware are scaled to the quadrillion-parameter regime, optical computers could have a $>8, 000\times$ energy-efficiency advantage over state-of-the-art digital-electronic processors that achieve 300 fJ/MAC.

Quantization

MoWE: Mixture of Weather Experts for Multiple Adverse Weather Removal

no code implementations24 Mar 2023 Yulin Luo, Rui Zhao, Xiaobao Wei, Jinwei Chen, Yijie Lu, Shenghao Xie, Tianyu Wang, Ruiqin Xiong, Ming Lu, Shanghang Zhang

Our MoWE achieves SOTA performance in upstream task on the proposed dataset and two public datasets, i. e. All-Weather and Rain/Fog-Cityscapes, and also have better perceptual results in downstream segmentation task compared to other methods.

Autonomous Driving Rain Removal

Learning Weather-General and Weather-Specific Features for Image Restoration Under Multiple Adverse Weather Conditions

no code implementations CVPR 2023 Yurui Zhu, Tianyu Wang, Xueyang Fu, Xuanyu Yang, Xin Guo, Jifeng Dai, Yu Qiao, Xiaowei Hu

Inspired by this observation, we design an efficient unified framework with a two-stage training strategy to explore the weather-general and weather-specific features.

Image Restoration

H2ONet: Hand-Occlusion-and-Orientation-Aware Network for Real-Time 3D Hand Mesh Reconstruction

no code implementations CVPR 2023 Hao Xu, Tianyu Wang, Xiao Tang, Chi-Wing Fu

First, we decouple hand mesh reconstruction into two branches, one to exploit finger-level non-occluded information and the other to exploit global hand orientation, with lightweight structures to promote real-time inference.

3D Hand Pose Estimation

From Random Search to Bandit Learning in Metric Measure Spaces

no code implementations19 May 2023 Chuying Han, Yasong Feng, Tianyu Wang

We show that, when the environment is noise-free, the output of random search converges to the optimal value in probability at rate $ \widetilde{\mathcal{O}} \left( \left( \frac{1}{T} \right)^{ \frac{1}{d_s} } \right) $, where $ d_s \ge 0 $ is the scattering dimension of the underlying function.

Hyperparameter Optimization

Optimizer's Information Criterion: Dissecting and Correcting Bias in Data-Driven Optimization

no code implementations16 Jun 2023 Garud Iyengar, Henry Lam, Tianyu Wang

We develop a general bias correction approach, building on what we call Optimizer's Information Criterion (OIC), that directly approximates the first-order bias and does not require solving any additional optimization problems.

Model Selection

Probabilistic and Semantic Descriptions of Image Manifolds and Their Applications

no code implementations6 Jul 2023 Peter Tu, Zhaoyuan Yang, Richard Hartley, Zhiwei Xu, Jing Zhang, Yiwei Fu, Dylan Campbell, Jaskirat Singh, Tianyu Wang

This paper begins with a description of methods for estimating image probability density functions that reflects the observation that such data is usually constrained to lie in restricted regions of the high-dimensional image space-not every pattern of pixels is an image.

Quantum-noise-limited optical neural networks operating at a few quanta per activation

no code implementations28 Jul 2023 Shi-Yuan Ma, Tianyu Wang, Jérémie Laydevant, Logan G. Wright, Peter L. McMahon

We experimentally demonstrated MNIST classification with a test accuracy of 98% using an optical neural network with a hidden layer operating in the single-photon regime; the optical energy used to perform the classification corresponds to 0. 008 photons per multiply-accumulate (MAC) operation, which is equivalent to 0. 003 attojoules of optical energy per MAC.

Image Classification

WALL-E: Embodied Robotic WAiter Load Lifting with Large Language Model

no code implementations30 Aug 2023 Tianyu Wang, YiFan Li, Haitao Lin, xiangyang xue, Yanwei Fu

The target instruction is then forwarded to a visual grounding system for object pose and size estimation, following which the robot grasps the object accordingly.

Language Modelling Large Language Model +3

Variational Inference for Scalable 3D Object-centric Learning

no code implementations25 Sep 2023 Tianyu Wang, Kee Siong Ng, Miaomiao Liu

We tackle the task of scalable unsupervised object-centric representation learning on 3D scenes.

Object Representation Learning +1

The hardware is the software

no code implementations20 Oct 2023 jeremie Laydevant, Logan G. Wright, Tianyu Wang, Peter L. McMahon

Human brains and bodies are not hardware running software: the hardware is the software.

Geometry-Calibrated DRO: Combating Over-Pessimism with Free Energy Implications

no code implementations8 Nov 2023 Jiashuo Liu, Jiayun Wu, Tianyu Wang, Hao Zou, Bo Li, Peng Cui

Machine learning algorithms minimizing average risk are susceptible to distributional shifts.

Cannot find the paper you are looking for? You can Submit a new open access paper.