Search

Most researchers focus on designing accurate crowd counting models with heavy parameters and computations but ignore the resource burden during the model deployment. A real-world scenario demands an efficient counting model with low-latency and high-performance. Knowledge distillation provides an elegant way to transfer knowledge from a complicated teacher model to a compact student model while maintaining accuracy. However, the student model receives the wrong guidance with the supervision of the teacher model due to the inaccurate information understood by the teacher in some cases. In this paper, we propose a dual-knowledge distillation (DKD) framework, which aims to reduce the side effects of the teacher model and transfer hierarchical knowledge to obtain a more efficient counting model. First, the student model is initialized with global information transferred by the teacher model via adaptive perspectives. Then, the self-knowledge distillation forces the student model to learn the knowledge by itself, based on intermediate feature maps and target map. Specifically, the optimal transport distance is utilized to measure the difference of feature maps between the teacher and the student to perform the distribution alignment of the counting area. Extensive experiments are conducted on four challenging datasets, demonstrating the superiority of DKD. When there are only approximately 6% of the parameters and computations from the original models, the student model achieves a faster and more accurate counting performance as the teacher model even surpasses it. [ABSTRACT FROM AUTHOR]

A compact MOSFET-C floating/grounded meminductor emulator (MIE) model is presented for high operating frequency and low power operation. The proposed MIE uses only 22 MOSFETs and two capacitors. Its performance is theoretically analyzed and rigorously verified using the Cadence Virtuoso software and hardware prototypes. The proposed MIE operates appropriately for a wide range of frequencies up to 5 MHz with $590 \mu \text{W}$ power consumption at a 180nm CMOS technology node and manifests important signature properties. The MIE layout area in 180 nm CMOS technology is $13107.5 \mu \text{m}^{2}$. To analyze the effects of statistical variations in MIE elements, extensive Monte Carlo simulations have been performed to demonstrate the robustness of the proposed MIE. For experimental validation, hardware prototypes have been developed and tested successfully. An MIE-based adaptive learning neuromorphic circuit is presented to show that it can mimic the behavioral responses of amoeba under varying environments such as temperature. [ABSTRACT FROM AUTHOR]

The fusion of human-centric design and artificial intelligence capabilities has opened up new possibilities for next-generation autonomous vehicles that go beyond traditional transportation. These vehicles can dynamically interact with passengers and adapt to their preferences. This article proposes a novel framework that leverages large language models (LLMs) to enhance the decision-making process in autonomous vehicles. By utilizing LLMs’ contextual understanding abilities with specialized tools, we aim to integrate the language and reasoning capabilities of LLMs into autonomous vehicles. Our research includes experiments in HighwayEnv, a collection of environments for autonomous driving and tactical decision-making tasks, to explore LLMs’ interpretation, interaction, and reasoning in various scenarios. We also examine some well-defined real-time personalized driving tasks, demonstrating how LLMs can influence driving behaviors based on drivers’ verbal commands. Our empirical results highlight the substantial advantages of utilizing chain-of-thought prompting, leading to improved driving decisions and showing the potential for LLMs to enhance personalized driving experiences through ongoing verbal feedback. The proposed framework aims to transform autonomous vehicle operations, offering personalized support, transparent decision making, and continuous learning to enhance safety and effectiveness. We achieve user-centric, transparent, and adaptive autonomous driving ecosystems supported by the integration of LLMs into autonomous vehicles. Experiment videos are available at https://youtube.com/playlist?list=PLgcRcf9w8BmLJi_fqTGq-7KCZsbpEIE4a&si=dhH9lgaeSmB5K94t. [ABSTRACT FROM AUTHOR]

Highly automated vehicles (HAVs) are exposed to various kinds of internal or external data faults, which may lead to disengagements and collisions. Detecting critical faults and evaluating the fault tolerance of HAVs, especially in terms of motion planning, is of great importance before HAVs are deployed on a large scale. However, due to the curse of dimensionality, which is led by high-dimensional fault features, current studies on fault tolerance testing have been limited to evaluating the vulnerability of a single location on motion planners. In this article, we propose a fault injection (FI) tool based on the idea of adaptive design of experiments (ADOE) to expediate the fault tolerance test of motion planners. The framework utilizes surrogate models (SMs) to highlight high-probability spaces where critical faults may appear and therefore narrow down the search space to accelerate the search process. ADOE helps strategically update SMs through the search process. Results show that the proposed ADOE-based FI method has great potential in accelerating fault tolerance testing for motion planners. Out of four evaluated SMs, support vector regression performs the best. We further train a classifier based on the multilayer perceptron (MLP) to judge whether a given fault in an unexamined scenario is critical or not. The F1 score of the MLP-based classifier reaches 90%. [ABSTRACT FROM AUTHOR]

Navigation or route guidance systems are designed to provide drivers with real-time travel information and the associated recommended routes for their trips. Classical route choice models typically rely on utility theory to represent drivers’ route choice behavior. Such choices, however, may not be optimal from both the individual and the system perspectives. This is simply due to the fact that drivers usually have imperfect knowledge about the time-varying traffic conditions. In this article, we explore and propose a new model-free deep reinforcement learning (DRL) approach to solving the adaptive route guidance problem based on microsimulation. The proposed approach consists of three interconnected algorithms, including a network edge labeling algorithm, a routing plan identification algorithm, and an adaptive route guidance algorithm. Simulation experiments on both a toy network and a real-world network of Suzhou, China, are performed to demonstrate the effectiveness of the proposed approach in terms of guiding a single vehicle as well as multiple vehicles through complex traffic environments. Comparative results confirm that the DRL approach outperforms the traditional shortest path method by further reducing the average travel time in the network. [ABSTRACT FROM AUTHOR]

Cross-domain aspect-based sentiment analysis has recently attracted significant attention, which can effectively alleviate the problem of lacking large-scale labeled data for supervised learning based methods. Most of current methods mainly focus on extracting domain-shared syntactic features to conduct the domain adaptation. Due to the language and syntax are diverse between domains, these methods lack generalization and even lead to syntactic transfer errors. External knowledge graphs have rich domain commonsense and share the relational structures between source and target domains. The domain-shared relational structure can effectively bridge the gap across domains and solve the problem of syntactic transfer errors. Moreover, not all the introduced external knowledge is equally important for the cross-domain aspect-based sentiment analysis. Motivated by these, we propose a knowledge-enhanced and topic-guided cross domain aspect-based sentiment analysis model with the domain-shared commonsense relational structure learning module and the topic-guided knowledge attention module. Extensive experiments are conducted and the experimental results evaluate the effectiveness of our proposed model. [ABSTRACT FROM AUTHOR]

Affective brain-computer interface based on electroencephalography (EEG) is an important branch in the field of affective computing. However, the individual differences in EEG emotional data and the noisy labeling problem in the subjective feedback seriously limit the effectiveness and generalizability of existing models. To tackle these two critical issues, we propose a novel transfer learning framework with Prototypical Representation based Pairwise Learning (PR-PL). The discriminative and generalized EEG features are learned for emotion revealing across individuals and the emotion recognition task is formulated as pairwise learning for improving the model tolerance to the noisy labels. More specifically, a prototypical learning is developed to encode the inherent emotion-related semantic structure of EEG data and align the individuals’ EEG features to a shared common feature space under consideration of the feature separability of both source and target domains. Based on the aligned feature representations, pairwise learning with an adaptive pseudo labeling method is introduced to encode the proximity relationships among samples and alleviate the label noises effect on modeling. Extensive results on two benchmark databases (SEED and SEED-IV) under four different cross-validation evaluation protocols validate the model reliability and stability across subjects and sessions. Compared to the literature, the average enhancement of emotion recognition across four different evaluation protocols is 2.04% (SEED) and 2.58% (SEED-IV). [ABSTRACT FROM AUTHOR]

Unsupervised domain adaptation (UDA) methods have become a research hotspot in remote sensing scene classification to reduce dependence on labeled samples. However, most current methods focus on extracting domain invariant features, ignoring the problem of large intraclass differences and the imbalanced sample numbers between categories in remote sensing images. To address these issues, we propose a remote sensing scene domain adaptive method based on middle-layer feature extraction and nuclear norm maximization (MFE-NM). In the MFE module, the middle-layer features of the feature extractor are randomly extracted and processed. Since the receptive field of the middle-layer features is smaller and the resolution is higher, the effective use of the middle-layer features can reduce the impact of image feature heterogeneity caused by large intraclass differences in remote sensing images. In addition, it can be concluded that the constrained nuclear norm can simultaneously improve the prediction diversity and discriminability of the model through theoretical derivation. Therefore, the NM module is proposed to solve the problem of reduced prediction diversity caused by entropy minimization methods when dealing with scene classification problems with imbalanced sample numbers between categories. Extensive experiments and analyses on three public remote sensing datasets demonstrate the effectiveness and competitiveness of our proposed method. [ABSTRACT FROM AUTHOR]

Contrastive learning has revolutionized the field of computer vision, learning rich representations from unlabeled data, which generalize well to diverse vision tasks. Consequently, it has become increasingly important to explain these approaches and understand their inner workings mechanisms. Given that contrastive models are trained with interdependent and interacting inputs and aim to learn invariance through data augmentation, the existing methods for explaining single-image systems (e.g., image classification models) are inadequate as they fail to account for these factors and typically assume independent inputs. Additionally, there is a lack of evaluation metrics designed to assess pairs of explanations, and no analytical studies have been conducted to investigate the effectiveness of different techniques used to explaining contrastive learning. In this work, we design visual explanation methods that contribute towards understanding similarity learning tasks from pairs of images. We further adapt existing metrics, used to evaluate visual explanations of image classification systems, to suit pairs of explanations and evaluate our proposed methods with these metrics. Finally, we present a thorough analysis of visual explainability methods for contrastive learning, establish their correlation with downstream tasks and demonstrate the potential of our approaches to investigate their merits and drawbacks. [ABSTRACT FROM AUTHOR]

Human-leading automated truck platooning has been an effective technique to improve traffic capacity and fuel economy and eliminate uncertainties of the traffic environment. Aiming for a tradeoff between the dynamic response of car following and energy-efficient platooning control, a predictive information multiagent soft actor–critic (PI-MASAC) control framework is proposed for a human-leading automated heavy-duty-truck platoon. In this framework, predictive information of environmental dynamics is modeled as the state representation of a deep reinforcement learning algorithm to address the uncertainties of a partially observable environment. In the truck model, the impact of intraplatoon aerodynamic interactions is modeled, which is used to design a constant spacing policy for platooning control. We demonstrate the effectiveness of our approach by testing the human-leading truck platoon under multiple scenarios compared to proximal policy optimization, an intelligent driver model, and linear-based cooperative adaptive cruise control. Our results show that the PI-MASAC learns a novel car-following strategy of peak shaving and valley filling and therefore significantly enhances energy savings by reducing high-intensity accelerations and decelerations. In addition, the PI-MASAC demonstrates its adaptability to various initial scenarios and exhibits good generalization to a larger platoon size. [ABSTRACT FROM AUTHOR]