arXiv subjects

Haotian Liu

Publications and source records attributed to Haotian Liu.

At least 91 records · Page 5Linked to original sources

arXiv2023

TMA: Temporal Motion Aggregation for Event-based Optical Flow

Event cameras have the ability to record continuous and detailed trajectories of objects with high temporal resolution, thereby providing intuitive motion cues for optical flow estimation. Nevertheless, most existing learning-based approaches for event optical flow estimation directly remould the paradigm of conventional images by representing the consecutive event stream as static frames, ignoring the inherent temporal continuity of event data. In this paper, we argue that temporal continuity is a vital element of event-based optical flow and propose a novel Temporal Motion Aggregation (TMA) approach to unlock its potential. Technically, TMA comprises three components: an event splitting strategy to incorporate intermediate motion information underlying the temporal context, a linear lookup strategy to align temporally fine-grained motion features and a novel motion pattern aggregation module to emphasize consistent patterns for motion feature enhancement. By incorporating temporally fine-grained motion information, TMA can derive better flow estimates than existing methods at early stages, which not only enables TMA to obtain more accurate final predictions, but also greatly reduces the demand for a number of refinements. Extensive experiments on DSEC-Flow and MVSEC datasets verify the effectiveness and superiority of our TMA. Remarkably, compared to E-RAFT, TMA achieves a 6\% improvement in accuracy and a 40\% reduction in inference time on DSEC-Flow. Code will be available at \url{https://github.com/ispc-lab/TMA}.

cs.CV↗

arXiv2023

Generate Anything Anywhere in Any Scene

Text-to-image diffusion models have attracted considerable interest due to their wide applicability across diverse fields. However, challenges persist in creating controllable models for personalized object generation. In this paper, we first identify the entanglement issues in existing personalized generative models, and then propose a straightforward and efficient data augmentation training strategy that guides the diffusion model to focus solely on object identity. By inserting the plug-and-play adapter layers from a pre-trained controllable diffusion model, our model obtains the ability to control the location and size of each generated personalized object. During inference, we propose a regionally-guided sampling technique to maintain the quality and fidelity of the generated images. Our method achieves comparable or superior fidelity for personalized objects, yielding a robust, versatile, and controllable text-to-image diffusion model that is capable of generating realistic and personalized images. Our approach demonstrates significant potential for various applications, such as those in art, entertainment, and advertising design.

cs.CV↗

arXiv2023

LLaVA-Med: Training a Large Language-and-Vision Assistant for Biomedicine in One Day

Conversational generative AI has demonstrated remarkable promise for empowering biomedical practitioners, but current investigations focus on unimodal text. Multimodal conversational AI has seen rapid progress by leveraging billions of image-text pairs from the public web, but such general-domain vision-language models still lack sophistication in understanding and conversing about biomedical images. In this paper, we propose a cost-efficient approach for training a vision-language conversational assistant that can answer open-ended research questions of biomedical images. The key idea is to leverage a large-scale, broad-coverage biomedical figure-caption dataset extracted from PubMed Central, use GPT-4 to self-instruct open-ended instruction-following data from the captions, and then fine-tune a large general-domain vision-language model using a novel curriculum learning method. Specifically, the model first learns to align biomedical vocabulary using the figure-caption pairs as is, then learns to master open-ended conversational semantics using GPT-4 generated instruction-following data, broadly mimicking how a layperson gradually acquires biomedical knowledge. This enables us to train a Large Language and Vision Assistant for BioMedicine (LLaVA-Med) in less than 15 hours (with eight A100s). LLaVA-Med exhibits excellent multimodal conversational capability and can follow open-ended instruction to assist with inquiries about a biomedical image. On three standard biomedical visual question answering datasets, LLaVA-Med outperforms previous supervised state-of-the-art on certain metrics. To facilitate biomedical multimodal research, we will release our instruction-following data and the LLaVA-Med model.

cs.CV↗

arXiv2023

GLIGEN: Open-Set Grounded Text-to-Image Generation

Large-scale text-to-image diffusion models have made amazing advances. However, the status quo is to use text input alone, which can impede controllability. In this work, we propose GLIGEN, Grounded-Language-to-Image Generation, a novel approach that builds upon and extends the functionality of existing pre-trained text-to-image diffusion models by enabling them to also be conditioned on grounding inputs. To preserve the vast concept knowledge of the pre-trained model, we freeze all of its weights and inject the grounding information into new trainable layers via a gated mechanism. Our model achieves open-world grounded text2img generation with caption and bounding box condition inputs, and the grounding ability generalizes well to novel spatial configurations and concepts. GLIGEN's zero-shot performance on COCO and LVIS outperforms that of existing supervised layout-to-image baselines by a large margin.

cs.CV↗

arXiv2023

Data-Efficient Image Quality Assessment with Attention-Panel Decoder

Blind Image Quality Assessment (BIQA) is a fundamental task in computer vision, which however remains unresolved due to the complex distortion conditions and diversified image contents. To confront this challenge, we in this paper propose a novel BIQA pipeline based on the Transformer architecture, which achieves an efficient quality-aware feature representation with much fewer data. More specifically, we consider the traditional fine-tuning in BIQA as an interpretation of the pre-trained model. In this way, we further introduce a Transformer decoder to refine the perceptual information of the CLS token from different perspectives. This enables our model to establish the quality-aware feature manifold efficiently while attaining a strong generalization capability. Meanwhile, inspired by the subjective evaluation behaviors of human, we introduce a novel attention panel mechanism, which improves the model performance and reduces the prediction uncertainty simultaneously. The proposed BIQA method maintains a lightweight design with only one layer of the decoder, yet extensive experiments on eight standard BIQA datasets (both synthetic and authentic) demonstrate its superior performance to the state-of-the-art BIQA methods, i.e., achieving the SRCC values of 0.875 (vs. 0.859 in LIVEC) and 0.980 (vs. 0.969 in LIVE).

cs.CV↗

arXiv2023

Learning Customized Visual Models with Retrieval-Augmented Knowledge

Image-text contrastive learning models such as CLIP have demonstrated strong task transfer ability. The high generality and usability of these visual models is achieved via a web-scale data collection process to ensure broad concept coverage, followed by expensive pre-training to feed all the knowledge into model weights. Alternatively, we propose REACT, REtrieval-Augmented CusTomization, a framework to acquire the relevant web knowledge to build customized visual models for target domains. We retrieve the most relevant image-text pairs (~3% of CLIP pre-training data) from the web-scale database as external knowledge, and propose to customize the model by only training new modualized blocks while freezing all the original weights. The effectiveness of REACT is demonstrated via extensive experiments on classification, retrieval, detection and segmentation tasks, including zero, few, and full-shot settings. Particularly, on the zero-shot classification task, compared with CLIP, it achieves up to 5.4% improvement on ImageNet and 3.7% on the ELEVATER benchmark (20 datasets).

cs.CV↗

arXiv2022

ELEVATER: A Benchmark and Toolkit for Evaluating Language-Augmented Visual Models

Learning visual representations from natural language supervision has recently shown great promise in a number of pioneering works. In general, these language-augmented visual models demonstrate strong transferability to a variety of datasets and tasks. However, it remains challenging to evaluate the transferablity of these models due to the lack of easy-to-use evaluation toolkits and public benchmarks. To tackle this, we build ELEVATER (Evaluation of Language-augmented Visual Task-level Transfer), the first benchmark and toolkit for evaluating(pre-trained) language-augmented visual models. ELEVATER is composed of three components. (i) Datasets. As downstream evaluation suites, it consists of 20 image classification datasets and 35 object detection datasets, each of which is augmented with external knowledge. (ii) Toolkit. An automatic hyper-parameter tuning toolkit is developed to facilitate model evaluation on downstream tasks. (iii) Metrics. A variety of evaluation metrics are used to measure sample-efficiency (zero-shot and few-shot) and parameter-efficiency (linear probing and full model fine-tuning). ELEVATER is a platform for Computer Vision in the Wild (CVinW), and is publicly released at at https://computer-vision-in-the-wild.github.io/ELEVATER/

cs.CV↗

arXiv2022

Reducing Action Space: Reference-Model-Assisted Deep Reinforcement Learning for Inverter-based Volt-Var Control

Reference-model-assisted deep reinforcement learning (DRL) for inverter-based Volt-Var Control (IB-VVC) in active distribution networks is proposed. We investigate that a large action space increases the learning difficulties of DRL and degrades the optimization performance in the process of generating data and training neural networks. To reduce the action space of DRL, we design a reference-model-assisted DRL approach. We introduce definitions of the reference model, reference-model-based optimization, and reference actions. The reference-model-assisted DRL learns the residual actions between the reference actions and optimal actions, rather than learning the optimal actions directly. Since the residual actions are considerably smaller than the optimal actions for a reference model, we can design a smaller action space for the reference-model-assisted DRL. It reduces the learning difficulties of DRL and optimises the performance of the reference-model-assisted DRL approach. It is noteworthy that the reference-model-assisted DRL approach is compatible with any policy gradient DRL algorithms for continuous action problems. This work takes the soft actor-critic algorithm as an example and designs a reference-model-assisted soft actor-critic algorithm. Simulations show that 1) large action space degrades the performance of DRL in the whole training stage, and 2) reference-model-assisted DRL requires fewer iteration times and returns a better optimization performance.

eess.SY↗

arXiv2022

Deflection and Gravitational lensing of null and timelike signals in the Kiselev black hole spacetime in the weak field limit

In this work we study the deflection and gravitational lensing of null and timelike signals in the Kiselev spacetime in the weak field limit, to investigate the effects of the equation of state parameter $ω$ and the matter amount parameter $α$. In doing this, we extend a perturbative method previously developed for asymptotically flat spacetimes whose metric functions have integer-power asymptotic expansions to the case that may or may not be asymptotically flat but with non-integer power expansions. It is found that in the asymptotically flat case ($-1/3<ω<0$) the deflection angles are expressable as quasi-power series of the dimensionless quantities $M/b,~b/r_{s,d}$ and $α/M^{1+3ω}$ where $M,~b,~r_{s,d}$ are respectively the lens mass, impact parameter and source/detector radius. A similar series exists for the non-asymptotically flat case of ($-1<ω<-1/3$), but with the closest radius $r_0$ replacing $b$. In the asymptotically flat (or non-flat) case, the increase of $α$ or decrease of $ω$ will increase (or increase) the deflection angle. Since the obtained deflection angles naturally take into account the finite distance effect of the source and the detector, we can establish an exact gravitational lensing equation, from which the apparent angles of the images and their magnifications are solved. It is found that generally for the asymptotically flat case, increasing $α$ or decreasing $ω$ will increase the apparent angles of the images. While for the non-asymptotically flat case, increasing $α$ or $ω$ will both lead to smaller apparent angles.

gr-qc↗

arXiv2022

Masked Discrimination for Self-Supervised Learning on Point Clouds

Masked autoencoding has achieved great success for self-supervised learning in the image and language domains. However, mask based pretraining has yet to show benefits for point cloud understanding, likely due to standard backbones like PointNet being unable to properly handle the training versus testing distribution mismatch introduced by masking during training. In this paper, we bridge this gap by proposing a discriminative mask pretraining Transformer framework, MaskPoint}, for point clouds. Our key idea is to represent the point cloud as discrete occupancy values (1 if part of the point cloud; 0 if not), and perform simple binary classification between masked object points and sampled noise points as the proxy task. In this way, our approach is robust to the point sampling variance in point clouds, and facilitates learning rich representations. We evaluate our pretrained models across several downstream tasks, including 3D shape classification, segmentation, and real-word object detection, and demonstrate state-of-the-art results while achieving a significant pretraining speedup (e.g., 4.1x on ScanNet) compared to the prior state-of-the-art Transformer baseline. Code is available at https://github.com/haotian-liu/MaskPoint.

cs.CV↗

arXiv2022

Reducing Learning Difficulties: One-Step Two-Critic Deep Reinforcement Learning for Inverter-based Volt-Var Control

A one-step two-critic deep reinforcement learning (OSTC-DRL) approach for inverter-based volt-var control (IB-VVC) in active distribution networks is proposed in this paper. Firstly, considering IB-VVC can be formulated as a single-period optimization problem, we formulate the IB-VVC as a one-step Markov decision process rather than the standard Markov decision process, which simplifies the DRL learning task. Then we design the one-step actor-critic DRL scheme which is a simplified version of recent DRL algorithms, and it avoids the issue of Q value overestimation successfully. Furthermore, considering two objectives of VVC: minimizing power loss and eliminating voltage violation, we utilize two critics to approximate the rewards of two objectives separately. It simplifies the approximation tasks of each critic, and avoids the interaction effect between two objectives in the learning process of critic. The OSTC-DRL approach integrates the one-step actor-critic DRL scheme and the two-critic technology. Based on the OSTC-DRL, we design two centralized DRL algorithms. Further, we extend the OSTC-DRL to multi-agent OSTC-DRL for decentralized IB-VVC and design two multi-agent DRL algorithms. Simulations demonstrate that the proposed OSTC-DRL has a faster convergence rate and a better control performance, and the multi-agent OSTC-DRL works well for decentralized IB-VVC problems.

cs.AI↗

arXiv2022

Controllable Production of Degenerate Fermi Gases of $^6$Li Atoms in the 2D-3D Crossover

The many-body physics in the dimensional crossover regime attracts much attention in cold atom experiments, but yet to explore systematically. One of the technical difficulties existed in the experiments is the lack of the experimental technique to quantitatively tune the atom occupation ratio of the different lattice bands. In this letter, we report such techniques in a process of transferring a 3D Fermi gas into a 1D optical lattice, where the capability of tuning the occupation of the energy band is realized by varying the trapping potentials of the optical dipole trap (ODT) and the lattice, respectively. We could tune a Fermi gas with the occupation in the lowest band from unity to 50$\%$ quantitatively. This provides a route to experimentally study the dependence of many-body interaction on the dimensionality in a Fermi gas.

cond-mat.quant-gas↗

arXiv2022

End-to-End Instance Edge Detection

Edge detection has long been an important problem in the field of computer vision. Previous works have explored category-agnostic or category-aware edge detection. In this paper, we explore edge detection in the context of object instances. Although object boundaries could be easily derived from segmentation masks, in practice, instance segmentation models are trained to maximize IoU to the ground-truth mask, which means that segmentation boundaries are not enforced to precisely align with ground-truth edge boundaries. Thus, the task of instance edge detection itself is different and critical. Since precise edge detection requires high resolution feature maps, we design a novel transformer architecture that efficiently combines a FPN and a transformer decoder to enable cross attention on multi-scale high resolution feature maps within a reasonable computation budget. Further, we propose a light weight dense prediction head that is applicable to both instance edge and mask detection. Finally, we use a penalty reduced focal loss to effectively train the model with point supervision on instance edges, which can reduce annotation costs. We demonstrate highly competitive instance edge detection performance compared to state-of-the-art baselines, and also show that the proposed task and loss are complementary to instance segmentation and object detection.

cs.CV↗

arXiv2021

M2MRF: Many-to-Many Reassembly of Features for Tiny Lesion Segmentation in Fundus Images

Feature reassembly is an essential component in modern CNN-based segmentation approaches, which includes feature downsampling and upsampling operators. Existing operators reassemble multiple features from a small predefined region into one for each target location independently. This may result in loss of spatial information, which could vanish activations caused by tiny lesions particularly when they cluster together. In this paper, we propose a many-to-many reassembly of features (M2MRF). It reassembles features in a dimension-reduced feature space and simultaneously aggregates multiple features inside a large predefined region into multiple target features. In this way, long range spatial dependencies are captured to maintain activations on tiny lesions. Experimental results on two lesion segmentation benchmarks, i.e. DDR and IDRiD, show that (1) our M2MRF outperforms existing feature reassembly operators; (2) equipped with our M2MRF, the HRNetv2 is able to achieve significant better performance to CNN-based segmentation methods and competitive even better performance to two recent transformer-based segmentation methods. Our code is made publicly available at https://github.com/CVIU-CSU/M2MRF-Lesion-Segmentation.

eess.IV↗

arXiv2021

Deflection and gravitational lensing of null and timelike signals in general asymptotically (anti-)de Sitter spacetimes

The deflection and gravitational lensing of light and massive particles in arbitrary static, spherically symmetric and asymptotically (anti-)de Sitter spacetimes are considered in this work. We first proved that for spacetimes whose metric satisfying certain conditions, the deflection of null rays with fixed closest distance will not depend on the cosmological constant $Λ$, while that of timelike signals and the apparent angle in gravitational lensing still depend on $Λ$. A two-step perturbative method is then developed to compute the change of the angular coordinate and total travel time in the weak field limit. The results are quasi-series of two small quantities, with the finite distance effect of the source/detector naturally taken into account. These results are verified by applying to some known de Sitter spacetimes. Using an exact gravitational lensing equation, we solved the apparent angles of the images and time delays between them and studied the effect of $Λ$ on them. It is found that generally, a small positive $Λ$ will decrease the apparent angle of images from both sides of the lens and increase the time delay between them. The time delay between signals from the same side of the lens but with different energy however, will be decreased by $Λ$.

gr-qc↗

arXiv2021

Cooling a Fermi gas with three-body recombination near a narrow Feshbach resonance

Three-body recombination is a phenomenon common in atomic and molecular collisions, producing heating in the system. However, we find the cooling effect of the three-body recombination of a 6Li Fermi gas near its s-wave narrow Feshbach resonance. Such counter-intuitive behavior is explained as follows, the threshold energy of the quasi-bounded Feshbach molecule acts as the knife of cooling, expelling the scattering atoms with selected kinetic energy from the trap. When the threshold energy happens to be larger than 3/2kBT, each lost atom in the three-body recombination process has more than 3kBT energy which results in cooling. The best cooling is found with the threshold energy set at about 3kBT, consistent with a theoretical model. The three-body recombination induced cooling raises potential applications for cooling complex atomic systems.

cond-mat.quant-gas↗

arXiv2021

Dual-Branch Network with Dual-Sampling Modulated Dice Loss for Hard Exudate Segmentation from Colour Fundus Images

Automated segmentation of hard exudates in colour fundus images is a challenge task due to issues of extreme class imbalance and enormous size variation. This paper aims to tackle these issues and proposes a dual-branch network with dual-sampling modulated Dice loss. It consists of two branches: large hard exudate biased learning branch and small hard exudate biased learning branch. Both of them are responsible for their own duty separately. Furthermore, we propose a dual-sampling modulated Dice loss for the training such that our proposed dual-branch network is able to segment hard exudates in different sizes. In detail, for the first branch, we use a uniform sampler to sample pixels from predicted segmentation mask for Dice loss calculation, which leads to this branch naturally be biased in favour of large hard exudates as Dice loss generates larger cost on misidentification of large hard exudates than small hard exudates. For the second branch, we use a re-balanced sampler to oversample hard exudate pixels and undersample background pixels for loss calculation. In this way, cost on misidentification of small hard exudates is enlarged, which enforces the parameters in the second branch fit small hard exudates well. Considering that large hard exudates are much easier to be correctly identified than small hard exudates, we propose an easy-to-difficult learning strategy by adaptively modulating the losses of two branches. We evaluate our proposed method on two public datasets and results demonstrate that ours achieves state-of-the-art performances.

cs.CV↗

arXiv2021

Bi-level Off-policy Reinforcement Learning for Volt/VAR Control Involving Continuous and Discrete Devices

In Volt/Var control (VVC) of active distribution networks(ADNs), both slow timescale discrete devices (STDDs) and fast timescale continuous devices (FTCDs) are involved. The STDDs such as on-load tap changers (OLTC) and FTCDs such as distributed generators should be coordinated in time sequence. Such VCC is formulated as a two-timescale optimization problem to jointly optimize FTCDs and STDDs in ADNs. Traditional optimization methods are heavily based on accurate models of the system, but sometimes impractical because of their unaffordable effort on modelling. In this paper, a novel bi-level off-policy reinforcement learning (RL) algorithm is proposed to solve this problem in a model-free manner. A Bi-level Markov decision process (BMDP) is defined to describe the two-timescale VVC problem and separate agents are set up for the slow and fast timescale sub-problems. For the fast timescale sub-problem, we adopt an off-policy RL method soft actor-critic with high sample efficiency. For the slow one, we develop an off-policy multi-discrete soft actor-critic (MDSAC) algorithm to address the curse of dimensionality with various STDDs. To mitigate the non-stationary issue existing the two agents' learning processes, we propose a multi-timescale off-policy correction (MTOPC) method by adopting importance sampling technique. Comprehensive numerical studies not only demonstrate that the proposed method can achieve stable and satisfactory optimization of both STDDs and FTCDs without any model information, but also support that the proposed method outperforms existing two-timescale VVC methods.

eess.SY↗