Fine-tuning has emerged as a popular technique in the field of transfer learning, demonstrating remarkable achievements in various data-scarce tasks. The performance of fine-tuning in deep convolutional neural networks depends on the selection of which parameters to fine-tune and freeze. However, it is difficult to determine which parameters in the pre-trained model need to be fine-tuned for a new task. This article proposes a filter-level discrete optimization model to identify the filter subset for fine-tuning, a core step of filter selection coding optimization. Due to the huge search space of the filter fine-tuning problem, we propose a filter interactivity decomposition strategy to find a valid search subspace (a smaller search subspace containing the optimal solution) by dividing the entire filter fine-tuning problem into multiple suboptimization problems. Based on the decomposition strategy, we design a microscale-searching transfer optimization algorithm, which solves each subproblem by searching the valid search subspace instead of the original search space of the filter fine-tuning problem. To verify the validity of the proposed algorithm, extensive experiments are conducted on seven publicly available image classification datasets: Stanford Dogs, MIT Indoors, Caltech 256-30, Caltech 256-60, Aircraft, UCF-101, and Omniglot. Experimental results show that the proposed method significantly improves the fine-tuning accuracy while effectively reducing the filter fine-tuning problem scale. Moreover, the proposed algorithm outperforms the state-of-the-art fine-tuning methods on the fine-tuning problem for transfer learning.

Abstract Deep neural networks are vulnerable to backdoor attacks. Previous backdoor attacks have mainly focused on images. Unlike images composed of regular pixels, 3D point clouds are composed of irregular three-dimensional XYZ coordinates, which are widely used in areas such as autonomous driving and 3D measurement. As many deep neural networks have been developed for processing 3D point clouds, these networks also face the risk of backdoor attacks. Nevertheless, backdoor attacks on 3D point clouds have rarely been investigated. This paper proposes a stealthy backdoor attack on point clouds in the physical world, aiming to generate trainable non-rigid deformations as backdoor patterns. Instead of directly adding backdoor patterns onto the point clouds, we deform the 3D space of the point clouds to a new space, ensuring that all point clouds have the same backdoor deformation. We use point cloud alignment to overcome the inconsistency of backdoor deformation caused by shifting and scaling in the physical world. We also propose a physical transformation layer to combat the physical transformations. Additionally, we propose mask contrast learning to eliminate pseudo backdoor patterns to make the network’s backdoor property stealthier. Extensive experiments indicate that the proposed method can achieve better attack success rates and stealthiness.

Research shows that deep neural networks are vulnerable to back-door attacks. The backdoor network behaves normally on clean examples, but once backdoor patterns are attached to examples, back-door examples will be classified into the target class. In the previous backdoor attack schemes, backdoor patterns are not stealthy and may be detected. Thus, to achieve the stealthiness of backdoor patterns, we explore an invisible and example-dependent backdoor attack scheme. Specifically, we employ the backdoor generation network to generate the invisible backdoor pattern for each example, and backdoor patterns are not generic to each other. However, without other measures, the backdoor attack scheme cannot bypass the neural cleanse detection. Thus, we propose adversarial training to bypass neural cleanse detection. Experiments show that the proposed backdoor attack achieves a considerable attack success rate, invisibility, and can bypass the existing defense strategies.

Different from supervised learning which requires all training examples to be labeled, Semi-Supervised Learning (SSL) learns from a few labeled training examples and a large number of unlabeled training examples. Recently, studies have shown that SSL is also vulnerable to backdoor attacks. However, their performance is poor. In this paper, we propose a novel unlabeled backdoor poisoning attack against SSL, where only poisoning unlabeled examples in the training set to inject the backdoor into the network. Specifically, our attack exploits the vulnerability of SSL algorithms in guessing pseudo labels of unlabeled examples. We propose a backdoor generation network to generate poisoned examples with both the backdoor property and misleading function, thus inducing the victim model itself to mislabel the poisoned examples as the target class and causing the backdoor to be injected. Our attack achieves favorable attack success rates on the SSL algorithm while bypassing backdoor defenses.

Underground square, as a recreation and activity place for citizens, represents the underground space quality of a city, so its spatial experience is of great significance. This research discussed the influence and effects of underground squares’ interface morphology on spatial experience. By field research of ten large-scale urban underground spaces, the basic elements and related numerical interval of underground squares’ interface morphology were summarized, and on basis of this, underground square virtual models with different interface morphology were built. Based on the platform of immersive virtual reality system (IVRS), combing the isovist method and the SD method, this paper made a quantitative analysis on the relevance between interface morphology and spatial experience, and the relevant compliance indicators of interface morphology were summarized. The experimental results show that the spatial experience is good at the interface density of 0.402 and interface opening’s aspect ratio of 2–3. This can provide reference and foundation for urban underground spatial design in future.

In this paper, a dilation of the LMI characterization is presented to address constrained robust model predictive control (MPC) for a class of uncertain linear systems with structured time-varying uncertainties. The uncertainty is described in linear fractional transformation (LFT) form. It is known such uncertain systems are popularly used in nonlinear system modeling and many other circumstances. By using parameter dependent Lyapunov functions, the designing conservativeness is reduced compared with some well-known MPC approaches. The proposed approach is applied to a two-mass-spring benchmark system to demonstrate the merits

It has been recognized that the dilation of the LMI characterizations has new potentials in dealing with such involved problems as multiobjective control, robust performance analysis or synthesis for real polytopic uncertainty and so on. In the model predictive control (MPC) area, Cuzzola et al. (2002) have proposed a technique which is based on the use of several Lyapunov functions each one corresponding to a different vertex of the uncertainty polytope. The main advantage of this approach compared to the other well-known techniques is the reduced conservativeness. However, this approach also increases the online computational demand, which partially limits its practicality. In this paper, an offline approach is proposed to reduce such online computational demand substantially. The approach is based on the concept of the asymptotically stable invariant ellipsoids and the closed-loop robust stability is guaranteed.