The rise of large language models (LLMs) such as ChatGPT presents new opportunities and challenges for computing education. While these models can provide instant assistance, their generality and lack of source grounding can produce inaccurate or unverifiable responses. This extended abstract introduces Alan, a course-specific virtual teaching assistant built on a retrieval-augmented generation (RAG) framework. Alan answers student questions using verified course resources such as lecture slides, lab handouts, notes, and recordings, and cites these sources directly in each response. Deployed in a large introductory programming course, Alan was used by over 250 students and logged more than 2,000 queries. A mixed-methods analysis combining usage data and survey responses shows that students found Alan most helpful for conceptual and administrative questions, valuing its transparency and course alignment. We conclude with lessons learned about integrating grounded AI support in computing classrooms.

Though caching on edge servers is widely acknowledged to be essential, it is not trivial to cache content on edge servers adaptively without any prior knowledge of the distribution of content popularity across the users. Several edge caching algorithms have been proposed in the literature based on multi-agent reinforcement learning (MARL) for dynamic control, however, they ignored the non-stationarity and partial-observability issues commonly existing in multi-agent systems. In an MARL-based edge caching application where agents collaborate towards a common goal, communication is essential as their decisions are jointly applied to improve collective intelligence. However, most existing methods proposed to exchange messages between agents have not considered the induced communication overhead, which is critical in practice with real-world multi-agent applications. In this paper, we propose a new MARL framework for edge caching where agents learn to construct, exchange and interpret collective messages for individual benefits, while controlling the complex collaborative task of cache replacement in a communication-efficient manner. With a standard edge caching model, we show that with limited communication and delays introduced, our proposed framework is able to outperform existing rule-based and learning-based caching policy alternatives.

With the growing interest in web services during the current COVID-19 outbreak, the demand for high-quality low-latency interactive applications has never been more apparent. Yet, packet losses are inevitable over the Internet, since it is based on UDP. In this paper, we propose Ivory, a new real-world system framework designed to support network adaptive error control in real-time communications, such as VoIP, using a recently proposed low-latency streaming code. We design and implement our prototype over UDP that can correct or retransmit lost packets conditional on network conditions and application requirements.To maintain the highest quality, Ivory attempts to correct as many lost packets as possible on-the-fly, yet incurring the smallest footprint in terms of coding overhead over the network. To achieve such an objective, Ivory uses a deep reinforcement learning agent to estimate the best coding parameters in real-time based on observed network states and experience learned. It learns offline the best coding parameters to use based on previously observed loss patterns and takes into account the round-trip time observed to decide on the optimum decoding delay for a low-latency application. Our extensive array of experiments shows that Ivory achieves a better trade-off between recovering packets and using lower redundancy than the state-of-the-art network adaptive streaming codes algorithms.

We introduce a novel network-adaptive algorithm that is suitable for alleviating network packet losses for low-latency interactive communications between a source and a destination. Our network-adaptive algorithm estimates in real-time the best parameters of a recently proposed streaming code that uses forward error correction (FEC) to correct both arbitrary and burst losses, which cause a crackling noise and undesirable jitters, respectively in audio. In particular, the destination estimates appropriate coding parameters based on its observed packet loss pattern and sends them back to the source for updating the underlying code. Besides, a new explicit construction of practical low-latency streaming codes that achieve the optimal tradeoff between the capability of correcting arbitrary losses and the capability of correcting burst losses is used. Simulation evaluations based on statistical losses and real-world packet loss traces reveal the following: (i) Our proposed network-adaptive algorithm combined with our optimal streaming codes can achieve significantly higher performance compared to uncoded and non-adaptive FEC schemes over UDP (User Datagram Protocol); (ii) Our explicit streaming codes can significantly outperform traditional MDS (maximum-distance separable) streaming schemes when they are used along with our network-adaptive algorithm. In addition, we study different factors that can affect the performance of our network-adaptive algorithm.

Traditional congestion control algorithms were designed with a hardwired heuristic mapping between packet- level events and predefined control actions in response to these events, and may fail to satisfy all the desirable performance goals as a result. In this paper, we seek to reconsider these fundamental goals in congestion control, and propose Eagle, a new congestion control algorithm to refine existing heuristics. Eagle takes advantage of expert knowledge from an existing algorithm, and uses deep reinforcement learning (DRL) to train a generalized model with the hope of learning from an expert. Learning by trial-and-error may not be as efficient as imitating a teacher; by the same token, DRL alone is not enough to guarantee good performance. In Eagle, we seek help from an expert congestion control algorithm, BBR, to help us train a long-short term memory (LSTM) neural network in the DRL agent, with the hope of making decisions that can be as good as or even better than the expert. With an extensive array of experiments, we discovered that Eagle is able to match and even outperform the performance of its teacher, and outperformed a large number of recent congestion control algorithms by a considerable margin.

We introduce a novel network-adaptive algorithm that is suitable for alleviating network packet losses for low-latency interactive communications between a source and a destination. Network packet losses happen in a bursty manner as well as an arbitrary manner, where the former is usually due to network congestion and the latter can be caused by unreliable wireless links. Our network-adaptive algorithm estimates in real time the best parameters of a recently proposed streaming code that corrects both arbitrary losses (which cause crackling noise in audio) and burst losses (which cause undesirable jitters and pauses in audio) using forward error correction (FEC). The network-adaptive algorithm updates the coding parameters in real time as follows: The destination estimates appropriate coding parameters based on its observed packet loss pattern and then the parameters are fed back to the source for updating the underlying code. In addition, a new explicit construction of practical low-latency streaming codes that achieve the optimal tradeoff between the capability of correcting arbitrary losses and the capability of correcting burst losses is provided. Simulation evaluations based on real-world packet loss traces reveal that our proposed network-adaptive algorithm combined with our optimal streaming codes achieves significantly higher reliability compared to uncoded and non-adaptive FEC schemes over UDP (User Datagram Protocol).

Vehicle-to-Everything (V2X) technology plays a critical role in maintaining road safety, avoiding accidents and controlling traffic flow. As self driving cars are expected to take over the roads, this paper discusses the intermediate phase in which manned and unmanned cars coexist. A heterogeneous network architecture that simultaneously serves manned and unmanned cars' different requirements in a suburban area is proposed and simulated using Riverbed Modeler. The feasibility of this architecture is examined in three different scenarios: Normal operation, congestion in both directions and Road Side Units (RSU) failure. In normal operation mode, traffic data is sent through Vehicle-to-Vehicle (V2V) and Vehicle-to-Infrastructure/Infrastructure-to-Vehicle (V2I/I2V or to RSU) using IEEE 802.11p and infotainment information is communicated as V2I/I2V using Long Term Evolution (LTE). A special case is highlighted and tested, in which congestion is in both directions. In such situation, data needs to be relayed to the nearest RSU using multi-hop communication. A fault-tolerant model is also proposed and analyzed in case of failure of RSU. The performance metrics are end-to-end delay, LTE response time, handover delay and packet loss ratio. The architecture proves its suitability by satisfying traffic control real time application requirements.