Hao Liu

Younggyo Seo, Danijar Hafner, Hao Liu et al. · deepmind, utoronto

Visual model-based reinforcement learning (RL) has the potential to enable sample-efficient robot learning from visual observations. Yet the current approaches typically train a single model end-to-end for learning both visual representations and dynamics, making it difficult to accurately model the interaction between robots and small objects. In this work, we introduce a visual model-based RL framework that decouples visual representation learning and dynamics learning. Specifically, we train an autoencoder with convolutional layers and vision transformers (ViT) to reconstruct pixels given masked convolutional features, and learn a latent dynamics model that operates on the representations from the autoencoder. Moreover, to encode task-relevant information, we introduce an auxiliary reward prediction objective for the autoencoder. We continually update both autoencoder and dynamics model using online samples collected from environment interaction. We demonstrate that our decoupling approach achieves state-of-the-art performance on a variety of visual robotic tasks from Meta-world and RLBench, e.g., we achieve 81.7% success rate on 50 visual robotic manipulation tasks from Meta-world, while the baseline achieves 67.9%. Code is available on the project website: https://sites.google.com/view/mwm-rl.

Kimin Lee, Hao Liu, Moonkyung Ryu et al.

Deep generative models have shown impressive results in text-to-image synthesis. However, current text-to-image models often generate images that are inadequately aligned with text prompts. We propose a fine-tuning method for aligning such models using human feedback, comprising three stages. First, we collect human feedback assessing model output alignment from a set of diverse text prompts. We then use the human-labeled image-text dataset to train a reward function that predicts human feedback. Lastly, the text-to-image model is fine-tuned by maximizing reward-weighted likelihood to improve image-text alignment. Our method generates objects with specified colors, counts and backgrounds more accurately than the pre-trained model. We also analyze several design choices and find that careful investigations on such design choices are important in balancing the alignment-fidelity tradeoffs. Our results demonstrate the potential for learning from human feedback to significantly improve text-to-image models.

Hao Liu, Xin Li, Mingming Gong et al. · tencent-ai

Recently, Table Structure Recognition (TSR) task, aiming at identifying table structure into machine readable formats, has received increasing interest in the community. While impressive success, most single table component-based methods can not perform well on unregularized table cases distracted by not only complicated inner structure but also exterior capture distortion. In this paper, we raise it as Complex TSR problem, where the performance degeneration of existing methods is attributable to their inefficient component usage and redundant post-processing. To mitigate it, we shift our perspective from table component extraction towards the efficient multiple components leverage, which awaits further exploration in the field. Specifically, we propose a seminal method, termed GrabTab, equipped with newly proposed Component Deliberator. Thanks to its progressive deliberation mechanism, our GrabTab can flexibly accommodate to most complex tables with reasonable components selected but without complicated post-processing involved. Quantitative experimental results on public benchmarks demonstrate that our method significantly outperforms the state-of-the-arts, especially under more challenging scenes.

Hao Liu, Alex Havrilla, Rongjie Lai et al. · gatech

Autoencoders have demonstrated remarkable success in learning low-dimensional latent features of high-dimensional data across various applications. Assuming that data are sampled near a low-dimensional manifold, we employ chart autoencoders, which encode data into low-dimensional latent features on a collection of charts, preserving the topology and geometry of the data manifold. Our paper establishes statistical guarantees on the generalization error of chart autoencoders, and we demonstrate their denoising capabilities by considering $n$ noisy training samples, along with their noise-free counterparts, on a $d$-dimensional manifold. By training autoencoders, we show that chart autoencoders can effectively denoise the input data with normal noise. We prove that, under proper network architectures, chart autoencoders achieve a squared generalization error in the order of $\displaystyle n^{-\frac{2}{d+2}}\log^4 n$, which depends on the intrinsic dimension of the manifold and only weakly depends on the ambient dimension and noise level. We further extend our theory on data with noise containing both normal and tangential components, where chart autoencoders still exhibit a denoising effect for the normal component. As a special case, our theory also applies to classical autoencoders, as long as the data manifold has a global parametrization. Our results provide a solid theoretical foundation for the effectiveness of autoencoders, which is further validated through several numerical experiments.

Hao Liu, Tom Zahavy, Volodymyr Mnih et al.

Large and diverse datasets have been the cornerstones of many impressive advancements in artificial intelligence. Intelligent creatures, however, learn by interacting with the environment, which changes the input sensory signals and the state of the environment. In this work, we aim to bring the best of both worlds and propose an algorithm that exhibits an exploratory behavior whilst it utilizes large diverse datasets. Our key idea is to leverage deep generative models that are pretrained on static datasets and introduce a dynamic model in the latent space. The transition dynamics simply mixes an action and a random sampled latent. It then applies an exponential moving average for temporal persistency, the resulting latent is decoded to image using pretrained generator. We then employ an unsupervised reinforcement learning algorithm to explore in this environment and perform unsupervised representation learning on the collected data. We further leverage the temporal information of this data to pair data points as a natural supervision for representation learning. Our experiments suggest that the learned representations can be successfully transferred to downstream tasks in both vision and reinforcement learning domains.

Haoyu Cao, Jiefeng Ma, Antai Guo et al.

Document Information Extraction (DIE) has attracted increasing attention due to its various advanced applications in the real world. Although recent literature has already achieved competitive results, these approaches usually fail when dealing with complex documents with noisy OCR results or mutative layouts. This paper proposes Generative Multi-modal Network (GMN) for real-world scenarios to address these problems, which is a robust multi-modal generation method without predefined label categories. With the carefully designed spatial encoder and modal-aware mask module, GMN can deal with complex documents that are hard to serialized into sequential order. Moreover, GMN tolerates errors in OCR results and requires no character-level annotation, which is vital because fine-grained annotation of numerous documents is laborious and even requires annotators with specialized domain knowledge. Extensive experiments show that GMN achieves new state-of-the-art performance on several public DIE datasets and surpasses other methods by a large margin, especially in realistic scenes.

Zhouzheng Li, Hao Liu

Beta-VAE is a very classical model for disentangled representation learning, the use of an expanding bottleneck that allow information into the decoder gradually is key to representation disentanglement as well as high-quality reconstruction. During recent experiments on such fascinating structure, we discovered that the total amount of latent variables can affect the representation learnt by the network: with very few latent variables, the network tend to learn the most important or principal variables, acting like a PCA; with very large numbers of latent variables, the variables tend to be more disentangled, and act like an ICA. Our assumption is that the competition between latent variables while trying to gain the most information bandwidth can lead to this phenomenon.

Xin Gao, Xueyuan Li, Hao Liu et al.

Platooning technology is renowned for its precise vehicle control, traffic flow optimization, and energy efficiency enhancement. However, in large-scale mixed platoons, vehicle heterogeneity and unpredictable traffic conditions lead to virtual bottlenecks. These bottlenecks result in reduced traffic throughput and increased energy consumption within the platoon. To address these challenges, we introduce a decision-making strategy based on nested graph reinforcement learning. This strategy improves collaborative decision-making, ensuring energy efficiency and alleviating congestion. We propose a theory of nested traffic graph representation that maps dynamic interactions between vehicles and platoons in non-Euclidean spaces. By incorporating spatio-temporal weighted graph into a multi-head attention mechanism, we further enhance the model's capacity to process both local and global data. Additionally, we have developed a nested graph reinforcement learning framework to enhance the self-iterative learning capabilities of platooning. Using the I-24 dataset, we designed and conducted comparative algorithm experiments, generalizability testing, and permeability ablation experiments, thereby validating the proposed strategy's effectiveness. Compared to the baseline, our strategy increases throughput by 10% and decreases energy use by 9%. Specifically, increasing the penetration rate of CAVs significantly enhances traffic throughput, though it also increases energy consumption.

Ying Fan, Olivia Watkins, Yuqing Du et al.

Learning from human feedback has been shown to improve text-to-image models. These techniques first learn a reward function that captures what humans care about in the task and then improve the models based on the learned reward function. Even though relatively simple approaches (e.g., rejection sampling based on reward scores) have been investigated, fine-tuning text-to-image models with the reward function remains challenging. In this work, we propose using online reinforcement learning (RL) to fine-tune text-to-image models. We focus on diffusion models, defining the fine-tuning task as an RL problem, and updating the pre-trained text-to-image diffusion models using policy gradient to maximize the feedback-trained reward. Our approach, coined DPOK, integrates policy optimization with KL regularization. We conduct an analysis of KL regularization for both RL fine-tuning and supervised fine-tuning. In our experiments, we show that DPOK is generally superior to supervised fine-tuning with respect to both image-text alignment and image quality. Our code is available at https://github.com/google-research/google-research/tree/master/dpok.

Hao Liu, Yang Yang, Yunxia Liu

Face Super-Resolution (FSR) aims to recover high-resolution (HR) face images from low-resolution (LR) ones. Despite the progress made by convolutional neural networks in FSR, the results of existing approaches are not ideal due to their low reconstruction efficiency and insufficient utilization of prior information. Considering that faces are highly structured objects, effectively leveraging facial priors to improve FSR results is a worthwhile endeavor. This paper proposes a novel network architecture called W-Net to address this challenge. W-Net leverages meticulously designed Parsing Block to fully exploit the resolution potential of LR image. We use this parsing map as an attention prior, effectively integrating information from both the parsing map and LR images. Simultaneously, we perform multiple fusions in various dimensions through the W-shaped network structure combined with the LPF(LR-Parsing Map Fusion Module). Additionally, we utilize a facial parsing graph as a mask, assigning different weights and loss functions to key facial areas to balance the performance of our reconstructed facial images between perceptual quality and pixel accuracy. We conducted extensive comparative experiments, not only limited to conventional facial super-resolution metrics but also extending to downstream tasks such as facial recognition and facial keypoint detection. The experiments demonstrate that W-Net exhibits outstanding performance in quantitative metrics, visual quality, and downstream tasks.

Zhe Gan, Liqun Chen, Weiyao Wang et al.

A Triangle Generative Adversarial Network ($Δ$-GAN) is developed for semi-supervised cross-domain joint distribution matching, where the training data consists of samples from each domain, and supervision of domain correspondence is provided by only a few paired samples. $Δ$-GAN consists of four neural networks, two generators and two discriminators. The generators are designed to learn the two-way conditional distributions between the two domains, while the discriminators implicitly define a ternary discriminative function, which is trained to distinguish real data pairs and two kinds of fake data pairs. The generators and discriminators are trained together using adversarial learning. Under mild assumptions, in theory the joint distributions characterized by the two generators concentrate to the data distribution. In experiments, three different kinds of domain pairs are considered, image-label, image-image and image-attribute pairs. Experiments on semi-supervised image classification, image-to-image translation and attribute-based image generation demonstrate the superiority of the proposed approach.