Low-Resolution Self-Attention for Semantic Segmentation

Semantic segmentation tasks naturally require high-resolution information for pixel-wise segmentation and global context information for class prediction. While existing vision transformers demonstrate promising performance, they often utilize high resolution context modeling, resulting in a computational bottleneck. In this work, we challenge conventional wisdom and introduce the Low-Resolution Self-Attention (LRSA) mechanism to capture global context at a significantly reduced computational cost. Our approach involves computing self-attention in a fixed low-resolution space regardless of the input image's resolution, with additional 3x3 depth-wise convolutions to capture fine details in the high-resolution space. We demonstrate the effectiveness of our LRSA approach by building the LRFormer, a vision transformer with an encoder-decoder structure. Extensive experiments on the ADE20K, COCO-Stuff, and Cityscapes datasets demonstrate that LRFormer outperforms state-of-the-art models. The code will be made available at https://github.com/yuhuan-wu/LRFormer.

Efficient Sharpness-aware Minimization for Improved Training of Neural Networks

Overparametrized Deep Neural Networks (DNNs) often achieve astounding performances, but may potentially result in severe generalization error. Recently, the relation between the sharpness of the loss landscape and the generalization error has been established by Foret et al. (2020), in which the Sharpness Aware Minimizer (SAM) was proposed to mitigate the degradation of the generalization. Unfortunately, SAM s computational cost is roughly double that of base optimizers, such as Stochastic Gradient Descent (SGD). This paper thus proposes Efficient Sharpness Aware Minimizer (ESAM), which boosts SAM s efficiency at no cost to its generalization performance. ESAM includes two novel and efficient training strategies-StochasticWeight Perturbation and Sharpness-Sensitive Data Selection. In the former, the sharpness measure is approximated by perturbing a stochastically chosen set of weights in each iteration; in the latter, the SAM loss is optimized using only a judiciously selected subset of data that is sensitive to the sharpness. We provide theoretical explanations as to why these strategies perform well. We also show, via extensive experiments on the CIFAR and ImageNet datasets, that ESAM enhances the efficiency over SAM from requiring 100% extra computations to 40% vis-a-vis base optimizers, while test accuracies are preserved or even improved.

Parallel Attention Network with Sequence Matching for Video Grounding

Given a video, video grounding aims to retrieve a temporal moment that semantically corresponds to a language query. In this work, we propose a Parallel Attention Network with Sequence matching (SeqPAN) to address the challenges in this task: multi-modal representation learning, and target moment boundary prediction. We design a self-guided parallel attention module to effectively capture self-modal contexts and cross-modal attentive information between video and text. Inspired by sequence labeling tasks in natural language processing, we split the ground truth moment into begin, inside, and end regions. We then propose a sequence matching strategy to guide start/end boundary predictions using region labels. Experimental results on three datasets show that SeqPAN is superior to state-of-the-art methods. Furthermore, the effectiveness of the self-guided parallel attention module and the sequence matching module is verified.

Video Corpus Moment Retrieval with Contrastive Learning

Given a collection of untrimmed and unsegmented videos, video corpus moment retrieval (VCMR) is to retrieve a temporal moment (i.e., a fraction of a video) that semantically corresponds to a given text query. As video and text are from two distinct feature spaces, there are two general approaches to address VCMR: (i) to separately encode each modality representations, then align the two modality representations for query processing, and (ii) to adopt fine-grained cross-modal interaction to learn multi-modal representations for query processing. While the second approach often leads to better retrieval accuracy, the first approach is far more efficient. In this paper, we propose a Retrieval and Localization Network with Contrastive Learning (ReLoCLNet) for VCMR. We adopt the first approach and introduce two contrastive learning objectives to refine video encoder and text encoder to learn video and text representations separately but with better alignment for VCMR. The video contrastive learning (VideoCL) is to maximize mutual information between query and candidate video at video-level. The frame contrastive learning (FrameCL) aims to highlight the moment region corresponds to the query at frame-level, within a video. Experimental results show that, although ReLoCLNet encodes text and video separately for efficiency, its retrieval accuracy is comparable with baselines adopting cross-modal interaction learning.

Natural Language Video Localization: A Revisit in Span-based Question Answering Framework

Natural Language Video Localization (NLVL) aims to locate a target moment from an untrimmed video that semantically corresponds to a text query. Existing approaches mainly solve the NLVL problem from the perspective of computer vision by formulating it as ranking, anchor, or regression tasks. These methods suffer from large performance degradation when localizing on long videos. In this work, we address the NLVL from a new perspective, i.e., span-based question answering (QA), by treating the input video as a text passage. We propose a video span localizing network (VSLNet), on top of the standard span-based QA framework (named VSLBase), to address NLVL. VSLNet tackles the differences between NLVL and span-based QA through a simple yet effective query-guided highlighting (QGH) strategy. QGH guides VSLNet to search for the matching video span within a highlighted region. To address the performance degradation on long videos, we further extend VSLNet to VSLNet-L by applying a multi-scale split-and-concatenation strategy. VSLNet-L first splits the untrimmed video into short clip segments; then, it predicts which clip segment contains the target moment and suppresses the importance of other segments. Finally, the clip segments are concatenated, with different confidences, to locate the target moment accurately. Extensive experiments on three benchmark datasets show that the proposed VSLNet and VSLNet-L outperform the state-of-the-art methods; VSLNet-L addresses the issue of performance degradation on long videos. Our study suggests that the span-based QA framework is an effective strategy to solve the NLVL problem.

Multiobjective Test Problems with Degenerate Pareto Fronts

In multiobjective optimization, a set of scalable test problems with a variety of features allows researchers to investigate and evaluate abilities of different optimization algorithms, and thus can help them to design and develop more effective and efficient approaches. Existing, commonly-used test problem suites are mainly focused on the situations where all the objectives are conflicting with each other. However, in some many-objective optimization problems, there may be unexpected characteristics among objectives, e.g., redundancy. This leads to a degenerate problem. In this paper, we systematically study degenerate problems. We abstract three generic characteristics of degenerate problems, and on the basis of these characteristics we present a set of test problems, in order to support the investigation of multiobjective search algorithms on problems with redundant objectives. To assess the proposed test problems, ten representative multiobjective evolutionary algorithms are tested. The results indicate that none of the tested algorithms is able to effectively solve these proposed problems, calling for the need of developing new approaches to addressing degenerate multi-objective problems.

Kernel Truncated Regression Representation for Robust Subspace Clustering

Subspace clustering aims to group data points into multiple clusters of which each corresponds to one subspace. Most existing subspace clustering methods assume that the data could be linearly represented with each other in the input space. In practice, however, this assumption is hard to be satisfied. To achieve nonlinear subspace clustering, we propose a novel method which consists of the following three steps: 1) projecting the data into a hidden space in which the data can be linearly reconstructed from each other; 2) calculating the globally linear reconstruction coefficients in the kernel space; 3) truncating the trivial coefficients to achieve robustness and block-diagonality, and then achieving clustering by solving a graph Laplacian problem. Our method has the advantages of a closed-form solution and capacity of clustering data points that lie in nonlinear subspaces. The first advantage makes our method efficient in handling large-scale data sets, and the second one enables the proposed method to address the nonlinear subspace clustering challenge. Extensive experiments on five real-world datasets demonstrate the effectiveness and the efficiency of the proposed method in comparison with ten state-of-the-art approaches regarding four evaluation metrics.

Locally linear representation for image clustering

It is a key to construct a similarity graph in graph-oriented subspace learning and clustering. In a similarity graph, each vertex denotes a data point and the edge weight represents the similarity between two points. There are two popular schemes to construct a similarity graph, i.e., pairwise distance based scheme and linear representation based scheme. Most existing works have only involved one of the above schemes and suffered from some limitations. Specifically, pairwise distance based methods are sensitive to the noises and outliers compared with linear representation based methods. On the other hand, there is the possibility that linear representation based algorithms wrongly select inter-subspaces points to represent a point, which will degrade the performance. In this paper, we propose an algorithm, called Locally Linear Representation (LLR), which integrates pairwise distance with linear representation together to address the problems. The proposed algorithm can automatically encode each data point over a set of points that not only could denote the objective point with less residual error, but also are close to the point in Euclidean space. The experimental results show that our approach is promising in subspace learning and subspace clustering.

How to Read Many-Objective Solution Sets in Parallel Coordinates

Rapid development of evolutionary algorithms in handling many-objective optimization problems requires viable methods of visualizing a high-dimensional solution set. Parallel coordinates which scale well to high-dimensional data are such a method, and have been frequently used in evolutionary many-objective optimization. However, the parallel coordinates plot is not as straightforward as the classic scatter plot to present the information contained in a solution set. In this paper, we make some observations of the parallel coordinates plot, in terms of comparing the quality of solution sets, understanding the shape and distribution of a solution set, and reflecting the relation between objectives. We hope that these observations could provide some guidelines as to the proper use of parallel coordinates in evolutionary many-objective optimization.