Oobleck enables resilient distributed training of large DNN models with
...
Generative tasks, such as text generation and question answering, hold a...
Lightness adaptation is vital to the success of image processing to avoi...
Image restoration (IR) has been an indispensable and challenging task in...
Calibration-based methods have dominated RAW image denoising under extre...
Key Information Extraction (KIE) is a challenging multimodal task that a...
Recent works have explored the fundamental role of depth estimation in
m...
Although previous co-speech gesture generation methods are able to synth...
Training visual reinforcement learning (RL) models in offline datasets i...
Large language models (LLMs) power a new generation of interactive AI
ap...
Image compression techniques typically focus on compressing rectangular
...
Image coding for machines (ICM) aims to compress images to support downs...
Video frame interpolation(VFI) has witnessed great progress in recent ye...
Although computational aesthetics evaluation has made certain achievemen...
3D representation disentanglement aims to identify, decompose, and manip...
Modern image inpainting systems, despite the significant progress, often...
Training deep neural networks (DNNs) is a major workload in datacenters
...
In this paper, we propose an embarrassingly simple yet highly effective
...
3D semantic scene completion (SSC) is an ill-posed task that requires
in...
The growth of pending legal cases in populous countries, such as India, ...
In recent years, we have witnessed the great advancement of Deep neural
...
Learned image compression has exhibited promising compression performanc...
Image forgery localization aims to identify forged regions by capturing
...
Few-shot image generation aims to generate data of an unseen category ba...
We present AIRS: Automatic Intrinsic Reward Shaping that intelligently a...
Computational aesthetics evaluation has made great achievements in the f...
Generating new fonts is a time-consuming and labor-intensive, especially...
We present MEM: Multi-view Exploration Maximization for tackling complex...
Large-scale vision-language models (VLMs) pre-trained on billion-level d...
Flow-guide synthesis provides a common framework for frame interpolation...
Exploration is critical for deep reinforcement learning in complex
envir...
In unsupervised domain adaptation (UDA), directly adapting from the sour...
Existing DNN serving solutions can provide tight latency SLOs while
main...
Time-lapse photography is employed in movies and promotional films becau...
Recently action recognition has received more and more attention for its...
In this paper, we present a ranking-based underwater image quality asses...
Aesthetic assessment of images can be categorized into two main forms:
n...
With the vigorous development of mobile photography technology, major mo...
Image aesthetic quality assessment is popular during the last decade. Be...
In recent years, image generation has made great strides in improving th...
Life-long learning aims at learning a sequence of tasks without forgetti...
Serverless computing is an emerging cloud computing paradigm that frees
...
Image Coding for Machines (ICM) aims to compress images for AI tasks ana...
With the continuous development of social software and multimedia techno...
This study uses TikTok (N = 8,173) to examine how short-form video platf...
This paper proposes Mandheling, the first system that enables highly
res...
Edge computing is a paradigm that shifts data processing services to the...
Existing general purpose frameworks for gigantic model training, i.e., m...
Adversarial learning has achieved remarkable performances for unsupervis...
In recent years, creative content generations like style transfer and ne...