With the popularity of the recent Transformer-based models represented by BERT, GPT-3 and ChatGPT, there has been state-of-the-art performance in a range of natural language processing tasks. However, the massive computations, huge memory footprint, and thus high latency of Transformer-based models is an inevitable challenge for the cloud with high real-time requirement. To tackle the issue, we propose BBCT, a method of block-wise bit-compression for transformer without retraining. Our method achieves more fine-grained compression of the whole transformer, including embedding, matrix multiplication, GELU, softmax, layer normalization, and all the intermediate results. As a case, we compress an efficient BERT with the method of BBCT. Our benchmark test results on General Language Understanding Evaluation (GLUE) show that BBCT can achieve less than 1% accuracy drop in most tasks.
We study contextual combinatorial bandits with probabilistically triggered arms (C$^2$MAB-T) under a variety of smoothness conditions that capture a wide range of applications, such as contextual cascading bandits and contextual influence maximization bandits. Under the triggering probability modulated (TPM) condition, we devise the C$^2$-UCB-T algorithm and propose a novel analysis that achieves an $\tilde{O}(d\sqrt{KT})$ regret bound, removing a potentially exponentially large factor $O(1/p_{\min})$, where $d$ is the dimension of contexts, $p_{\min}$ is the minimum positive probability that any arm can be triggered, and batch-size $K$ is the maximum number of arms that can be triggered per round. Under the variance modulated (VM) or triggering probability and variance modulated (TPVM) conditions, we propose a new variance-adaptive algorithm VAC$^2$-UCB and derive a regret bound $\tilde{O}(d\sqrt{T})$, which is independent of the batch-size $K$. As a valuable by-product, we find our analysis technique and variance-adaptive algorithm can be applied to the CMAB-T and C$^2$MAB~setting, improving existing results there as well. We also include experiments that demonstrate the improved performance of our algorithms compared with benchmark algorithms on synthetic and real-world datasets.
Parametric optimization is an important product design technique, especially in the context of the modern parametric feature-based CAD paradigm. Realizing its full potential, however, requires a closed loop between CAD and CAE (i.e., CAD/CAE integration) with automatic design modifications and simulation updates. Conventionally the approach of model conversion is often employed to form the loop, but this way of working is hard to automate and requires manual inputs. As a result, the overall optimization process is too laborious to be acceptable. To address this issue, a new method for parametric optimization is introduced in this paper, based on a unified model representation scheme called eXtended Voxels (XVoxels). This scheme hybridizes feature models and voxel models into a new concept of semantic voxels, where the voxel part is responsible for FEM solving, and the semantic part is responsible for high-level information to capture both design and simulation intents. As such, it can establish a direct mapping between design models and analysis models, which in turn enables automatic updates on simulation results for design modifications, and vice versa -- effectively a closed loop between CAD and CAE. In addition, robust and efficient geometric algorithms for manipulating XVoxel models and efficient numerical methods (based on the recent finite cell method) for simulating XVoxel models are provided. The presented method has been validated by a series of case studies of increasing complexity to demonstrate its effectiveness. In particular, a computational efficiency improvement of up to 55.8 times the existing FCM method has been seen.
While features of different scales are perceptually important to visual inputs, existing vision transformers do not yet take advantage of them explicitly. To this end, we first propose a cross-scale vision transformer, CrossFormer. It introduces a cross-scale embedding layer (CEL) and a long-short distance attention (LSDA). On the one hand, CEL blends each token with multiple patches of different scales, providing the self-attention module itself with cross-scale features. On the other hand, LSDA splits the self-attention module into a short-distance one and a long-distance counterpart, which not only reduces the computational burden but also keeps both small-scale and large-scale features in the tokens. Moreover, through experiments on CrossFormer, we observe another two issues that affect vision transformers' performance, i.e. the enlarging self-attention maps and amplitude explosion. Thus, we further propose a progressive group size (PGS) paradigm and an amplitude cooling layer (ACL) to alleviate the two issues, respectively. The CrossFormer incorporating with PGS and ACL is called CrossFormer++. Extensive experiments show that CrossFormer++ outperforms the other vision transformers on image classification, object detection, instance segmentation, and semantic segmentation tasks. The code will be available at: https://github.com/cheerss/CrossFormer.
Pseudo-Labeling has emerged as a simple yet effective technique for semi-supervised object detection (SSOD). However, the inevitable noise problem in pseudo-labels significantly degrades the performance of SSOD methods. Recent advances effectively alleviate the classification noise in SSOD, while the localization noise which is a non-negligible part of SSOD is not well-addressed. In this paper, we analyse the localization noise from the generation and learning phases, and propose two strategies, namely pseudo-label correction and noise-unaware learning. For pseudo-label correction, we introduce a multi-round refining method and a multi-vote weighting method. The former iteratively refines the pseudo boxes to improve the stability of predictions, while the latter smoothly self-corrects pseudo boxes by weighing the scores of surrounding jittered boxes. For noise-unaware learning, we introduce a loss weight function that is negatively correlated with the Intersection over Union (IoU) in the regression task, which pulls the predicted boxes closer to the object and improves localization accuracy. Our proposed method, Pseudo-label Correction and Learning (PCL), is extensively evaluated on the MS COCO and PASCAL VOC benchmarks. On MS COCO, PCL outperforms the supervised baseline by 12.16, 12.11, and 9.57 mAP and the recent SOTA (SoftTeacher) by 3.90, 2.54, and 2.43 mAP under 1\%, 5\%, and 10\% labeling ratios, respectively. On PASCAL VOC, PCL improves the supervised baseline by 5.64 mAP and the recent SOTA (Unbiased Teacherv2) by 1.04 mAP on AP$^{50}$.
Recent advancements toward perception and decision-making of flexible endoscopes have shown great potential in computer-aided surgical interventions. However, owing to modeling uncertainty and inter-patient anatomical variation in flexible endoscopy, the challenge remains for efficient and safe navigation in patient-specific scenarios. This paper presents a novel data-driven framework with self-contained visual-shape fusion for autonomous intelligent navigation of flexible endoscopes requiring no priori knowledge of system models and global environments. A learning-based adaptive visual servoing controller is proposed to online update the eye-in-hand vision-motor configuration and steer the endoscope, which is guided by monocular depth estimation via a vision transformer (ViT). To prevent unnecessary and excessive interactions with surrounding anatomy, an energy-motivated shape planning algorithm is introduced through entire endoscope 3-D proprioception from embedded fiber Bragg grating (FBG) sensors. Furthermore, a model predictive control (MPC) strategy is developed to minimize the elastic potential energy flow and simultaneously optimize the steering policy. Dedicated navigation experiments on a robotic-assisted flexible endoscope with an FBG fiber in several phantom environments demonstrate the effectiveness and adaptability of the proposed framework.
Video-Text Retrieval (VTR) aims to search for the most relevant video related to the semantics in a given sentence, and vice versa. In general, this retrieval task is composed of four successive steps: video and textual feature representation extraction, feature embedding and matching, and objective functions. In the last, a list of samples retrieved from the dataset is ranked based on their matching similarities to the query. In recent years, significant and flourishing progress has been achieved by deep learning techniques, however, VTR is still a challenging task due to the problems like how to learn an efficient spatial-temporal video feature and how to narrow the cross-modal gap. In this survey, we review and summarize over 100 research papers related to VTR, demonstrate state-of-the-art performance on several commonly benchmarked datasets, and discuss potential challenges and directions, with the expectation to provide some insights for researchers in the field of video-text retrieval.
A recent line of works apply machine learning techniques to assist or rebuild cost-based query optimizers in DBMS. While exhibiting superiority in some benchmarks, their deficiencies, e.g., unstable performance, high training cost, and slow model updating, stem from the inherent hardness of predicting the cost or latency of execution plans using machine learning models. In this paper, we introduce a learning-to-rank query optimizer, called Lero, which builds on top of a native query optimizer and continuously learns to improve the optimization performance. The key observation is that the relative order or rank of plans, rather than the exact cost or latency, is sufficient for query optimization. Lero employs a pairwise approach to train a classifier to compare any two plans and tell which one is better. Such a binary classification task is much easier than the regression task to predict the cost or latency, in terms of model efficiency and accuracy. Rather than building a learned optimizer from scratch, Lero is designed to leverage decades of wisdom of databases and improve the native query optimizer. With its non-intrusive design, Lero can be implemented on top of any existing DBMS with minimal integration efforts. We implement Lero and demonstrate its outstanding performance using PostgreSQL. In our experiments, Lero achieves near optimal performance on several benchmarks. It reduces the plan execution time of the native optimizer in PostgreSQL by up to 70% and other learned query optimizers by up to 37%. Meanwhile, Lero continuously learns and automatically adapts to query workloads and changes in data.
Data-driven materials design often encounters challenges where systems require or possess qualitative (categorical) information. Metal-organic frameworks (MOFs) are an example of such material systems. The representation of MOFs through different building blocks makes it a challenge for designers to incorporate qualitative information into design optimization. Furthermore, the large number of potential building blocks leads to a combinatorial challenge, with millions of possible MOFs that could be explored through time consuming physics-based approaches. In this work, we integrated Latent Variable Gaussian Process (LVGP) and Multi-Objective Batch-Bayesian Optimization (MOBBO) to identify top-performing MOFs adaptively, autonomously, and efficiently without any human intervention. Our approach provides three main advantages: (i) no specific physical descriptors are required and only building blocks that construct the MOFs are used in global optimization through qualitative representations, (ii) the method is application and property independent, and (iii) the latent variable approach provides an interpretable model of qualitative building blocks with physical justification. To demonstrate the effectiveness of our method, we considered a design space with more than 47,000 MOF candidates. By searching only ~1% of the design space, LVGP-MOBBO was able to identify all MOFs on the Pareto front and more than 97% of the 50 top-performing designs for the CO$_2$ working capacity and CO$_2$/N$_2$ selectivity properties. Finally, we compared our approach with the Random Forest algorithm and demonstrated its efficiency, interpretability, and robustness.