arXiv subjects

Bai Liu

Publications and source records attributed to Bai Liu.

5 recordsLinked to original sources

arXiv2024

LLM4GEN: Leveraging Semantic Representation of LLMs for Text-to-Image Generation

Diffusion models have exhibited substantial success in text-to-image generation. However, they often encounter challenges when dealing with complex and dense prompts involving multiple objects, attribute binding, and long descriptions. In this paper, we propose a novel framework called \textbf{LLM4GEN}, which enhances the semantic understanding of text-to-image diffusion models by leveraging the representation of Large Language Models (LLMs). It can be seamlessly incorporated into various diffusion models as a plug-and-play component. A specially designed Cross-Adapter Module (CAM) integrates the original text features of text-to-image models with LLM features, thereby enhancing text-to-image generation. Additionally, to facilitate and correct entity-attribute relationships in text prompts, we develop an entity-guided regularization loss to further improve generation performance. We also introduce DensePrompts, which contains $7,000$ dense prompts to provide a comprehensive evaluation for the text-to-image generation task. Experiments indicate that LLM4GEN significantly improves the semantic alignment of SD1.5 and SDXL, demonstrating increases of 9.69\% and 12.90\% in color on T2I-CompBench, respectively. Moreover, it surpasses existing models in terms of sample quality, image-text alignment, and human evaluation.

cs.CV↗

arXiv2022

Universal Policy Tracking: Scheduling for Wireless Networks with Delayed State Observation

Numerous scheduling algorithms have been proposed to optimize various performance metrics like throughput, delay and utility in wireless networks. However, these algorithms often require instantaneous access to network state information, which is not always available. While network stability can sometimes be achieved with delayed state information, other performance metrics such as latency may degrade. Thus, instead of simply stabilizing the system, our goal is to design a framework that can mimic arbitrary scheduling algorithms with performance guarantees. A naive approach is to make decisions directly with delayed information, but we show that such methods may lead to poor performance. Instead, we propose the Universal Tracking (UT) algorithm that can mimic the actions of arbitrary scheduling algorithms under observation delay. We rigorously show that the performance gap between UT and the scheduling algorithm being tracked is bounded by constants. Our numerical experiments show that UT significantly outperforms the naive approach in various applications.

cs.PF↗

arXiv2021

Easy and Efficient Transformer : Scalable Inference Solution For large NLP model

Recently, large-scale transformer-based models have been proven to be effective over various tasks across many domains. Nevertheless, applying them in industrial production requires tedious and heavy works to reduce inference costs. To fill such a gap, we introduce a scalable inference solution: Easy and Efficient Transformer (EET), including a series of transformer inference optimization at the algorithm and implementation levels. First, we design highly optimized kernels for long inputs and large hidden sizes. Second, we propose a flexible CUDA memory manager to reduce the memory footprint when deploying a large model. Compared with the state-of-the-art transformer inference library (Faster Transformer v4.0), EET can achieve an average of 1.40-4.20x speedup on the transformer decoder layer with an A100 GPU

cs.CL↗

arXiv2020

RL-QN: A Reinforcement Learning Framework for Optimal Control of Queueing Systems

With the rapid advance of information technology, network systems have become increasingly complex and hence the underlying system dynamics are often unknown or difficult to characterize. Finding a good network control policy is of significant importance to achieve desirable network performance (e.g., high throughput or low delay). In this work, we consider using model-based reinforcement learning (RL) to learn the optimal control policy for queueing networks so that the average job delay (or equivalently the average queue backlog) is minimized. Traditional approaches in RL, however, cannot handle the unbounded state spaces of the network control problem. To overcome this difficulty, we propose a new algorithm, called Reinforcement Learning for Queueing Networks (RL-QN), which applies model-based RL methods over a finite subset of the state space, while applying a known stabilizing policy for the rest of the states. We establish that the average queue backlog under RL-QN with an appropriately constructed subset can be arbitrarily close to the optimal result. We evaluate RL-QN in dynamic server allocation, routing and switching problems. Simulation results show that RL-QN minimizes the average queue backlog effectively.

cs.PF↗

arXiv2016

Global optimization framework for real-time route guidance via variable message sign

Variable message sign (VMS) is an effective traffic management tool for congestion mitigation. The VMS is primarily used as a means of providing factual travel information or genuine route guidance to travelers. However, this may be rendered sub-optimal on a network level by potential network paradoxes and lack of consideration for its cascading effect on the rest of the network. This paper focuses on the design of optimal display strategy of VMS in response to real-time traffic information and its coordination with other intelligent transportation systems such as signal control, in order to explore the full potential of real-time route guidance in combating congestion. We invoke the linear decision rule framework to design the optimal on-line VMS strategy, and test its effectiveness in conjunction with on-line signal control. A simulation case study is conducted on a real-world test network in China, which shows the advantage of the proposed adaptive VMS display strategy over genuine route guidance, as well as its synergies with on-line signal control for congestion mitigation.

math.OC↗