Delivering an immersive experience to virtual reality (VR) users through wireless connectivity offers the freedom to engage from anywhere at any time. Nevertheless, it is challenging to ensure seamless wireless connectivity that delivers real-time and high-quality videos to the VR users. This paper proposes a field of view (FoV) aware caching for mobile edge computing (MEC)-enabled wireless VR network. In particular, the FoV of each VR user is cached/prefetched at the base stations (BSs) based on the caching strategies tailored to each BS. Specifically, decentralized and personalized federated learning (DP-FL) based caching strategies with guarantees are presented. Considering VR systems composed of multiple VR devices and BSs, a DP-FL caching algorithm is implemented at each BS to personalize content delivery for VR users. The utilized DP-FL algorithm guarantees a probably approximately correct (PAC) bound on the conditional average cache hit. Further, to reduce the cost of communicating gradients, one-bit quantization of the stochastic gradient descent (OBSGD) is proposed, and a convergence guarantee of $\mathcal{O}(1/\sqrt{T})$ is obtained for the proposed algorithm, where $T$ is the number of iterations. Additionally, to better account for the wireless channel dynamics, the FoVs are grouped into multicast or unicast groups based on the number of requesting VR users. The performance of the proposed DP-FL algorithm is validated through realistic VR head-tracking dataset, and the proposed algorithm is shown to have better performance in terms of average delay and cache hit as compared to baseline algorithms.

LLMs such as ChatGPT have been widely adopted by students in higher education as tools for learning programming and related concepts. However, it remains unclear how effective students are and what strategies students use while learning with LLMs. Since the majority of students' experiences in online self-learning have come through using search engines such as Google, evaluating AI tools in this context can help us address these gaps. In this mixed methods research, we conducted an exploratory within-subjects study to understand how CS2 students learn programming concepts using both LLMs as well as traditional online methods such as educational websites and videos to examine how students approach learning within and across both scenarios. We discovered that students found it easier to learn a more difficult concept using traditional methods than using ChatGPT. We also found that students ask fewer follow-ups and use more keyword-based queries for search engines while their prompts to LLMs tend to explicitly ask for information.

There has been a plethora of recent and historical work on this topic, finding different ways to help networks alleviate the issue of catastrophic forgetting --- where a network trained on tasks A_0 through A_i, forgets these to differing degrees when trained on tasks A_i 1 onward. Most methods can be divided into regularisation based, memory based or meta-learning based. One relatively recent work is GEM (gradient of episodic memory) (and relatedly A-GEM). This works by storing examples from seen tasks in an episodic memory. When learning a new task, the gradient update is modified such that it does not increase the loss on examples from previous tasks (these are represented by the examples in memory).

The paper introduces a clear, simple generalisation of two established continual learning methods (GEM and A-GEM) which performs very well in a thorough empirical evaluation. All reviewers and the AC value the effort that the authors put in their response. There is consensus that the work has merit and all reviewers recommend accepting the paper (R1 and R4 raised their score).

This paper studies the problem of online classification with partial feedback under the new constraint that the policy satisfies a fairness (equality of false positives) constraint at each round. The paper leverages careful modification of a number of technical tools to prove the O(sqrt(T)) regret with gamma O(T (-1/4)) fairness rate. In particular, they reduce the partial feedback setting to a contextual bandits problem, construct an approximate "fair oracle" using a modification of the reductions approach to fair classification, and then modify ILOVETOCONBANDITS to use this approximate oracle. The relevant inspiration is clearly cited, and the main contribution is combining these tools to effectively handle the fairness constraint in the online learning problem. The proposed algorithm is intuitive: accept everyone in the early rounds to gather data and use this data to determine which classifiers satisfy the constrain.

The authors cast the task of parallel training as a learning problem, allowing data driven decisions to be made instead of the hand-crafted rules. The topic is relevant and the results are impactful. The comprehensive ablation studies performed to evaluate the system are also appreciated. Several aspects of the proposed system have room for improvement, both in terms of scope and quality. However, that doesn't seem to be a crucial problem with the paper but rather room for follow up works.

The paper is well written in general although it contains mistakes and ignores some related work. In particular, it is not clear whether the corollaries (whose proofs are given in the appendix) are sold as contributions or not. Many of their implications are already known in the machine learning literature (see details below). Mistakes: - Wrong curvature (minor mistake): The hyperboloid defined in Eq. (3) is said to have a constant curvature of -1/K 2 in the submission. As explained in detail in Section 3.4 of Chapter 3 of the second edition of [1A] (or also in the following references [1C] and [1D]), its curvature is actually -1/K.

In lines 245-248 the authors discuss a fair comparison between the different methods and mention their effort to keep the total number of coupling layers the same between several methods the same. Can the authors please also comment on the difference in the number of parameters? As the coupling layers in M-Flows don't always act on data of the same dimensionality as regular AF flows, the number of parameters can be different, even with the same number of coupling layers. For the celebA dataset, have you tried to train M-Flows with different n then 512? 4. Can you explain in the main text on a high level why including the SCANDAL loss consistently leads to a larger closure for all methods (lower closure is better). In general, since the supplementary material contains so much more material, it would help the reader if you refer more frequently to the relevant parts of the supplementary material in the main text.

All reviewers agree that the presented technique for simultaneous manifold and density estimation is interesting and novel. However, they also agree that the paper leaves important questions open. While one of the reviewers would like to see a stronger statistical analysis before acceptance, the others believe that the paper is above acceptance threshold and that the community would benefit from its communication. To address the concerns of the reviewers, the camera-ready paper needed to include at least the following results: 1. Include results that investigate if the invertible nature of the normalising flow in the decoder is useful by e.g considering a version of the Me-flow where g is not constrained to be invertible. In the same vein, a comparison with a simple VAE baseline should be included. Investigate how the results on CelebA depend on the latent dimension n.

Humming away in offices on Capitol Hill, in the Pentagon and in the White House is a technology that represents the pragmatism, efficiency and unsentimental nature of American bureaucracy: the autopen. It is a device that stores a person's signature, replicating it as needed using a mechanical arm that holds a real pen. Like many technologies, this rudimentary robotic signature-maker has always provoked ambivalence. We invest signatures with meaning, particularly when the signer is well known. During the George W Bush administration, the secretary of defence, Donald Rumsfeld, generated a small wave of outrage when reporters revealed that he had been using an autopen for his signature on the condolence letters that he sent to the families of fallen soldiers. Fans of singer Bob Dylan expressed ire when they discovered that the limited edition of his book The Philosophy of Modern Song, which cost nearly 600 and came with an official certificate "attesting to its having been individually signed by Dylan", in fact had made unlimited use of an autopen. Dylan took the unusual step of issuing a statement on his Facebook page: "With contractual deadlines looming," Dylan wrote, "the idea of using an autopen was suggested to me, along with the assurance that this kind of thing is done'all the time' in the art and literary worlds."