Search Results for author:
Found 26 papers, 8 papers with code
Revisiting Optimal Convergence Rate for Smooth and Non-convex Stochastic Decentralized Optimization
While (Lu and Sa, 2021) have recently provided an optimal rate for non-convex stochastic decentralized optimization with weight matrices defined over linear graphs, the optimal rate with general weight matrices remains unclear.
Recurrent Dynamic Embedding for Video Object Segmentation
In this paper, we propose a Recurrent Dynamic Embedding (RDE) to build a memory bank of constant size.
Ranked #16 on
Semi-Supervised Video Object Segmentation
on MOSE
Multi-View Consistent Generative Adversarial Networks for 3D-aware Image Synthesis
To address this challenge, we propose Multi-View Consistent Generative Adversarial Networks (MVCGAN) for high-quality 3D-aware image synthesis with geometry constraints.
RCL: Recurrent Continuous Localization for Temporal Action Detection
Temporal representation is the cornerstone of modern action detection techniques.
Disentangled Representation Learning for Text-Video Retrieval
Cross-modality interaction is a critical component in Text-Video Retrieval (TVR), yet there has been little examination of how different influencing factors for computing interaction affect performance.
Ranked #10 on
Video Retrieval
on MSR-VTT-1kA
(using extra training data)
SEEG: Semantic Energized Co-Speech Gesture Generation
Talking gesture generation is a practical yet challenging task which aims to synthesize gestures in line with speech.
Ranked #6 on
Gesture Generation
on TED Gesture Dataset
Exponential Graph is Provably Efficient for Decentralized Deep Training
Experimental results on a variety of tasks and models demonstrate that decentralized (momentum) SGD over exponential graphs promises both fast and high-quality training.
Communicate Then Adapt: An Effective Decentralized Adaptive Method for Deep Training
Decentralized adaptive gradient methods, in which each node averages only with its neighbors, are critical to save communication and wall-clock training time in deep learning tasks.
Communication Efficient SGD via Gradient Sampling With Bayes Prior
Different from all of them, we regard large and small gradients selection as the exploitation and exploration of gradient information, respectively.
Accelerating Gossip SGD with Periodic Global Averaging
Communication overhead hinders the scalability of large-scale distributed training.
OR-Net: Pointwise Relational Inference for Data Completion under Partial Observation
Specifically, we expect to approximate the real joint distribution over the partial observation and latent variables, thus infer the unseen targets respectively.
DecentLaM: Decentralized Momentum SGD for Large-batch Deep Training
Experimental results on a variety of computer vision tasks and models demonstrate that DecentLaM promises both efficient and high-quality training.
Learning Position and Target Consistency for Memory-based Video Object Segmentation
To address this limitation, we propose to Learn position and target Consistency framework for Memory-based video object segmentation, termed as LCM.
Multiple Object Tracking with Correlation Learning
Recent works have shown that convolutional networks have substantially improved the performance of multiple object tracking by simultaneously learning detection and appearance features.
Few-Shot Incremental Learning with Continually Evolved Classifiers
First, we adopt a simple but effective decoupled learning strategy of representations and classifiers that only the classifiers are updated in each incremental session, which avoids knowledge forgetting in the representations.
Ranked #7 on
Few-Shot Class-Incremental Learning
on CIFAR-100
Self-supervised Video Representation Learning by Context and Motion Decoupling
A key challenge in self-supervised video representation learning is how to effectively capture motion information besides context bias.
Virtual ID Discovery from E-commerce Media at Alibaba: Exploiting Richness of User Click Behavior for Visual Search Relevance
Benefiting from exploration of user click data, our networks are more effective to encode richer supervision and better distinguish real-shot images in terms of category and feature.
Large Scale Long-tailed Product Recognition System at Alibaba
In this paper, we present a novel side information based large scale visual recognition co-training~(SICoT) system to deal with the long tail problem by leveraging the image related side information.
Distribution Adaptive INT8 Quantization for Training CNNs
Researches have demonstrated that low bit-width (e. g., INT8) quantization can be employed to accelerate the inference process.
Large-Scale Visual Search with Binary Distributed Graph at Alibaba
For a deployed visual search system with several billions of online images in total, building a billion-scale offline graph in hours is essential, which is almost unachievable by most existing methods.
Train a One-Million-Way Instance Classifier for Unsupervised Visual Representation Learning
However, scaling up the classification task from thousands of semantic labels to millions of instance labels brings specific challenges including 1) the large-scale softmax computation; 2) the slow convergence due to the infrequent visiting of instance samples; and 3) the massive number of negative classes that can be noisy.
Fashion Focus: Multi-modal Retrieval System for Video Commodity Localization in E-commerce
Nowadays, live-stream and short video shopping in E-commerce have grown exponentially.
Visual Search at Alibaba
We hope visual search at Alibaba becomes more widely incorporated into today's commercial applications.
Large-Scale Training System for 100-Million Classification at Alibaba
In the last decades, extreme classification has become an essential topic for deep learning.
Weakly Supervised Learning with Side Information for Noisy Labeled Images
In many real-world datasets, like WebVision, the performance of DNN based classifier is often limited by the noisy labeled data.
Deep Reinforcement Learning of Volume-guided Progressive View Inpainting for 3D Point Scene Completion from a Single Depth Image
Given a single depth image, our method first goes through the 3D volume branch to obtain a volumetric scene reconstruction as a guide to the next view inpainting step, which attempts to make up the missing information; the third step involves projecting the volume under the same view of the input, concatenating them to complete the current view depth, and integrating all depth into the point cloud.
