This research examines the impact of ChatGPT on computer science education, focusing on its application in learning programming languages like SQL, C++, Python, and C#. Through a survey of 149 university students, the study identifies both the advantages and challenges of using ChatGPT as a virtual lab assistant. The results show that ChatGPT offers considerable assistance to students, especially in providing prompt feedback, facilitating debugging, and clarifying complex programming concepts. However, the research also points out significant challenges, including the potential for over-reliance on AI tools and worries about the accuracy of the responses generated by AI. Several students reported experiencing misleading or incomplete information from ChatGPT, indicating a need for enhancements in its accuracy and dependability. To address these challenges, the study proposes a transition from assignments focused on theory to personalized, project-based activities that foster independent problem-solving. In summary, this research sheds light on the advantages and obstacles of incorporating ChatGPT into computer science courses. Although ChatGPT provides beneficial support for learning, its function should be supplementary rather than central to programming education. The results highlight the necessity of thoughtful integration, promoting self-directed learning while utilizing AI's capabilities to foster engagement and offer immediate assistance.

In many disciplines, the growth of online courses was propelled by the COVID-19 pandemic, but this trend moderated as health concerns receded. This study presents a comparative analysis of students' and faculty's perceptions towards online labs in three computing-related disciplines, a year after the pandemic. Through a survey with 769 students and 20 faculty responses, we found students were overwhelmingly positive about their online lab experience - as were faculty. Students and instructors both agree that 1) the perception of online lab courses has improved since the COVID-19 pandemic, and 2) it is crucial to continue investing in technology infrastructure to enhance the quality and accessibility of both online and in-person labs. However, students and instructors disagree on two issues: 1) teamwork for lab activities and assignments (i.e., faculty tended to have a more optimistic view of online collaborative activities); and 2) modality for lab sessions (i.e., student preferences were evenly split between synchronous and asynchronous labs while faculty mostly preferred synchronous online labs). In this lightning talk, we will also discuss best practices worth promoting despite the added effort for faculty.

For airborne radar, it is an urgent task to detect moving targets in clutter background. At present, automatic feature extraction is an effective way to improve the performance of moving target detection (MTD). In this letter, a probability of false alarm (PFA)-controllable detection method based on attention-enhanced convolutional neural network (CNN) with multiscale depth-separable convolution (MSDCAN) is proposed. First, the airborne radar model is established to obtain the radar space-time echo dataset, and then, the trained CNN is used to extract the features of the radar space-time echo data. The designed CNN can automatically learn the different features between the target and the clutter and give the corresponding probability. Finally, the false alarm controllable (FAC) module is used as a discriminator to divide the feature vector into two categories, so as to achieve the control of the PFA. Experimental results demonstrate that compared to other CNN-based detectors, the proposed detector exhibits a higher detection probability on both simulated dataset and the dataset containing real data.

To distill more information from training data, more parameters are introduced into machine learning models. As a result, communication becomes the bottleneck of Distributed Machine Learning (DML) systems. To alleviate the communication resource contention among DML jobs, which prolongs the time to train machine learning models, in machine learning clusters, JointPS is proposed in this paper. JointPS first minimizes the completion time of a single training epoch for each DML job via jointly optimizing the parameter server placement and flow scheduling, and predicts the number of remaining training epochs for each DML job by leveraging a dynamic model fitting method. Then, JointPS can estimate the remaining time to complete each DML job. According to such estimation, JointPS schedules DML jobs following the Minimum Remaining Time First (MRTF) principle to minimize the average job completion time. To the best of our knowledge, JointPS should be the first work that minimizes the average completion time of network-intensive DML training jobs by jointly optimizing the parameter server placement and flow scheduling without modifying the DML models and training procedures. Through both testbed experiments and extensive simulations, we demonstrate that JointPS can reduce the average completion time of DML jobs by up to 88% compared with state-of-the-art technology.

This paper presents the development of a new learning platform in Virtual Reality to create a more immersive and intuitive learning experience for introduction of programming courses at an intermediate level. This platform is designed to create a central hub for interactive courseware and facilitate distance learning in our post COVID world. Utilizing Virtual Reality, the application teaches specific topics in Computer Science using scripted animations, tutorials, and interactive games. A pilot study was conducted to evaluate the user experience and learning outcomes. Participants of this study reported they were more engaged and motivated in learning programing concepts. We found the virtual learning modules helped to explain abstract concepts and provided better hands-on experiences.

In this paper, we propose EdgePS, an advanced parameter server approach for distributed machine learning in edge computing scenarios. Different from the Conventional Parameter Server (CPS) approach, which performs parameter aggregation after every local training epoch, EdgePS synchronizes the parameters of all workers only when the local training cannot improve the global model performance. We first analyze how the local training will impact the performance of the global model, and then design algorithms to determine when the best time is to perform the parameter aggregation. Both real testbed experiments and extensive large scale simulations demonstrate that EdgePS can train a practical machine learning model, e.g., VGG-16, with up to 59.28% less time compared with the CPS approach. With the same training time, EdgePS can improve model accuracy by up to 30.19 % compared with the state-of-the-art distributed machine learning algorithm designed for edge computing scenarios.

Riders of public transportation in recent years tend to expect the intelligence of transit systems to progress in tandem with modern technology, with a key factor of this being an improved on-time prediction and scheduling of bus systems. This work explores various regression algorithms, including Support Vector Regression and Linear Regression, along with evaluation of loss functions and regulation terms to address the bus arrival time problem. The experiment is a case study of Route 280 from the Foothill Transit Agency, and resulted in Huber Regression with ElasticNet penalty giving the most accurate prediction.

Universities across the U.S. have moved to various virtual teaching models in response to the health threats caused by the COVID-19 pandemic. When converting to an online-only mode, STEM and related fields face additional challenges over the lab portion of courses because laboratory courses and active-learning projects frequently require specialized equipment and manual dexterity interactions. In this paper, we report the results of a study on students' perceptions about online learning during the initial phase of the pandemic at a public university in California, U.S. We focus on the overall reaction to the rapid conversion to online, the negative impressions created, “structural” concerns that would be difficult to mitigate, concerns readily amenable to mitigation, and side effects such as impact on equity. Twenty-five recommendations for those factors deemed improvable are provided.

This paper investigates how to estimate instantaneous fuel consumption using smartphones and OBD-II (On-board Diagnostics) adapters. Although most of the new cars have instant miles per gallon readout feature, the readings from those dashboard displays are usually proprietary and are difficult to record. Not to mention older cars do not have this feature. In this paper, we describe a system and associated algorithms to monitor fuel consumption of gasoline-powered vehicles in real time at second-level granularity. Specifically, we propose two algorithms: 1) Powertrain-based Model, which is derived from estimating an engine's fuel injection rate, and 2) Vehicle Dynamics-based Model, which considers fuel consumption in terms of the mechanical work applied to a vehicle. They are designed for vehicles with and without OBD-II adaptors respectively. The proposed system is compatible with most of the passenger vehicles and can be easily deployed. We evaluate our system in a field test and show that it can successfully estimate instantaneous fuel consumption, the average difference between estimation results and the ground truth is about 6%.

Recent advances in connected vehicle technologies enable vehicles and signal controllers to cooperate and improve the traffic management at intersections. This paper explores the opportunity for cooperative and integrated vehicle and intersection control for energy efficiency (CIVIC-E2) to contribute to a more sustainable transportation system. We propose a two-level approach that jointly optimizes the traffic signal timing and vehicles' approach speed, with the objective being to minimize total energy consumption for all vehicles passing through an isolated intersection. More specifically, at the intersection level, a dynamic programming algorithm is designed to find the optimal signal timing by explicitly considering the arrival time and energy profile of each vehicle. At the vehicle level, a model predictive control strategy is adopted to ensure that vehicles pass through the intersection in a timely fashion. Our simulation study has shown that the proposed CIVIC-E2system can significantly improve intersection performance under various traffic conditions. Compared with conventional fixed-time and actuated signal control strategies, the proposed algorithm can reduce energy consumption and queue length by up to 31% and 95%, respectively.

Location information is playing a significant role in nowadays mobile applications. Performing indoor localization with existing WiFi infrastructure and smartphone motion sensor through statistical filtering has been proven to be a feasible solution. Many literature have resorted to particle filters to deal with the multi-modal and non-Gaussian problem associated with the filtering process. Although grid-based filters can approximate the true densities better compared with particle filters, their computational cost is extremely expensive, especially for large areas. In this paper, we develop a novel asymmetric grid-based filter to accommodate both high-resolution requirement and computational cost-efficiency. The evaluation over an indoor area of 3750m2has shown that our proposed method can achieve a median error of only 2.72m, which is 0.17m more accurate with only 24% of the computation cost compared to particle filters.

This paper investigates a new application of vehicular sensing: detecting and reporting the slippery road conditions. We describe a system and associated algorithm to monitor vehicle skidding events using smartphones and OBD-II (On board Diagnostics) adapters. This system, which we call the VehSense, gathers data from smartphone inertial sensors and vehicle wheel speed sensors, and processes the data to monitor slippery road conditions in real-time. Specifically, two speed readings are collected: 1) ground speed, which is estimated by vehicle acceleration and rotation, and 2) wheel speed, which is retrieved from the OBD-II interface. The mismatch between these two speeds is used to infer a skidding event. Without tapping into vehicle manufactures' proprietary data (e.g., antilock braking system), VehSense is compatible with most of the passenger vehicles, and thus can be easily deployed. We evaluate our system on snow-covered roads at Buffalo, and show that it can detect vehicle skidding effectively.

Different from the conventional operation mode in existing taxi dispatch systems, in this paper, we envision a new cyber-technology enabled taxi dispatch system which can efficiently provide vacant taxis with cruising route suggestions, not to respond to any specific pick-up request but instead, hoping to find prospective customers (such system is also complementary to the conventional operation mode). We address the Taxi Cruising Guidance (TCG) problem with the objective being to minimize the Global Vacant Rate (GVR), which is defined as the ratio of traveling miles with no passenger onboard, to the total traveling miles in a given time period. We propose a number of heuristic solutions and conduct comprehensive performance evaluations based on large-scale simulations. A case study is also presented by utilizing real traces collected from taxis in the city of Shanghai. As part of our research, we leverage a well-known microscopic traffic simulator (called TRANSIMS) to demonstrate that the application of TCG is also beneficial to traffic management.

We consider a promising application in Vehicular Cyber-Physical Systems (VCPS) called On-road Ad Delivery (OAD), where targeted advertisements are delivered via roadside APs to attract commuters to nearby shops. Different from most existing works on VANETs which only focused on a single technical area, this work on OAD involves technical elements from human factors, cyber systems and transportation systems since a commuter's shopping decision depends on e.g. the attractiveness of the ads, the induced detour, and traffic conditions on different routes. In this paper, we address a new optimization problem in OAD whose goal is to schedule ad messages and allocate a limited amount of AP bandwidth so as to maximize the system-wide performance in terms of total realized utilities (TRU) of the delivered ads. A number of efficient heuristics are proposed to deal with ad message scheduling and AP bandwidth allocation. Besides largescale simulations, we also present a case study in a more realistic scenario utilizing real traces collected from taxis in the city of Shanghai. In addition, we use a commercial traffic simulator (PARAMICS) to show that our proposed solutions are also useful for traffic management in terms of balancing vehicular traffic and alleviating congestion.

Current transfer-incapable carpooling (TIC) scheme cannot fully utilize vehicles' available space because a carpooling passenger has to go from her origin to her destination by getting a ride from only one vehicle. This is akin to insist on delivering some packets only using one-hop communications, which usually performs worse than allowing multi-hop communications. In this paper, inspired by the “Store-and-Forward” strategy used in Delay-Tolerant Networks (DTN), we propose a new carpooling paradigm called transfer-allowed carpooling (TAC), with which each passenger can be served by more than one vehicle to go from her origin to her destination, thus increasing the carpooling performance. In particular, when given a) a number of carpooling requests (each with a maximum waiting-time and a maximum number of transfers for a passenger), and b) a list of participating vehicles (each specifying a maximum detour distance for a driver), we address a new optimization problem called Transfer-Allowed Carpooling whose objective is to maximize the successful carpooling ratio (SCR). Two effective strategies have been proposed from a driver and passenger standpoint, respectively. In addition to conducting large-scale simulations, we also present a case study in a more realistic setting by utilizing real routes collected from taxis in the city of Shanghai. Our major results are: 1) the proposed TAC approach can significantly improve SCR (by 35% to 60%), compared to the traditional TIC approach; and 2) allowing one transfer (i.e., the maximum number of transfers=1) improves the carpooling efficiency most, while allowing more than one transfer does not bring any noticeable benefits.