Search

Assessing and enhancing human learning through question-answering is vital, yet automating this process remains challenging. While large language models (LLMs) excel at summarization and query responses, their ability to generate meaningful questions for learners is underexplored. We propose Savaal, a scalable question-generation system with three objectives: (i) scalability, enabling question generation from hundreds of pages of text (ii) depth of understanding, producing questions beyond factual recall to test conceptual reasoning, and (iii) domain-independence, automatically generating questions across diverse knowledge areas. Instead of providing an LLM with large documents as context, Savaal improves results with a three-stage processing pipeline. Our evaluation with 76 human experts on 71 papers and PhD dissertations shows that Savaal generates questions that better test depth of understanding by 6.5X for dissertations and 1.5X for papers compared to a direct-prompting LLM baseline. Notably, as document length increases, Savaal's advantages in higher question quality and lower cost become more pronounced., Comment: Kimia Noorbakhsh, Joseph Chandler, and Pantea Karimi contributed equally to the work

We prove that if a given reaction network $\mathcal{N}$ has a weakly reversible deficiency zero realization for all choice of rate constants, then there exists a $\textit{unique}$ weakly reversible deficiency zero network $\mathcal{N}'$ such that $\mathcal{N}$ is realizable by $\mathcal{N}'$. Additionally, we propose an algorithm to find this weakly reversible deficiency zero network $\mathcal{N}'$ when it exists., Comment: 21 pages, 3 figures

Cloud platforms remain underutilized despite multiple proposals to improve their utilization (e.g., disaggregation, harvesting, and oversubscription). Our characterization of the resource utilization of virtual machines (VMs) in Azure reveals that, while CPU is the main underutilized resource, we need to provide a solution to manage all resources holistically. We also observe that many VMs exhibit complementary temporal patterns, which can be leveraged to improve the oversubscription of underutilized resources. Based on these insights, we propose Coach: a system that exploits temporal patterns for all-resource oversubscription in cloud platforms. Coach uses long-term predictions and an efficient VM scheduling policy to exploit temporally complementary patterns. We introduce a new general-purpose VM type, called CoachVM, where we partition each resource allocation into a guaranteed and an oversubscribed portion. Coach monitors the oversubscribed resources to detect contention and mitigate any potential performance degradation. We focus on memory management, which is particularly challenging due to memory's sensitivity to contention and the overhead required to reassign it between CoachVMs. Our experiments show that Coach enables platforms to host up to ~26% more VMs with minimal performance degradation., Comment: To appear in 30th ACM International Conference on Architectural Support for Programming Languages and Operating Systems, Volume 2 (ASPLOS'25). 15 pages

Compute Express Link (CXL) is widely-supported interconnect standard that promises to enable memory disaggregation in data centers. CXL allows for memory pooling, which can be used to create a shared memory space across multiple servers. However, CXL does not specify how to actually build a memory pool. Existing proposals for CXL memory pools are expensive, as they require CXL switches or large multi-headed devices. In this paper, we propose a new design for CXL memory pools that is cost-effective. We call these designs Octopus topologies. Our design uses small CXL devices that can be made cheaply and offer fast access latencies. Specifically, we propose asymmetric CXL topologies where hosts connect to different sets of CXL devices. This enables pooling and sharing memory across multiple hosts even as each individual CXL device is only connected to a small number of hosts. Importantly, this uses hardware that is readily available today. We also show the trade-off in terms of CXL pod size and cost overhead per host. Octopus improves the Pareto frontier defined by prior policies, e.g., offering to connect 3x as many hosts at 17% lower cost per host.

Many problems that cloud operators solve are computationally expensive, and operators often use heuristic algorithms (that are faster and scale better than optimal) to solve them more efficiently. Heuristic analyzers enable operators to find when and by how much their heuristics underperform. However, these tools do not provide enough detail for operators to mitigate the heuristic's impact in practice: they only discover a single input instance that causes the heuristic to underperform (and not the full set), and they do not explain why. We propose XPlain, a tool that extends these analyzers and helps operators understand when and why their heuristics underperform. We present promising initial results that show such an extension is viable.

Computational wave imaging (CWI) extracts hidden structure and physical properties of a volume of material by analyzing wave signals that traverse that volume. Applications include seismic exploration of the Earth's subsurface, acoustic imaging and non-destructive testing in material science, and ultrasound computed tomography in medicine. Current approaches for solving CWI problems can be divided into two categories: those rooted in traditional physics, and those based on deep learning. Physics-based methods stand out for their ability to provide high-resolution and quantitatively accurate estimates of acoustic properties within the medium. However, they can be computationally intensive and are susceptible to ill-posedness and nonconvexity typical of CWI problems. Machine learning-based computational methods have recently emerged, offering a different perspective to address these challenges. Diverse scientific communities have independently pursued the integration of deep learning in CWI. This review delves into how contemporary scientific machine-learning (ML) techniques, and deep neural networks in particular, have been harnessed to tackle CWI problems. We present a structured framework that consolidates existing research spanning multiple domains, including computational imaging, wave physics, and data science. This study concludes with important lessons learned from existing ML-based methods and identifies technical hurdles and emerging trends through a systematic analysis of the extensive literature on this topic., Comment: 29 pages, 11 figures

Graph neural networks (GNNs) are powerful graph-based machine-learning models that are popular in various domains, e.g., social media, transportation, and drug discovery. However, owing to complex data representations, GNNs do not easily allow for human-intelligible explanations of their predictions, which can decrease trust in them as well as deter any collaboration opportunities between the AI expert and non-technical, domain expert. Here, we first discuss the two papers that aim to provide GNN explanations to domain experts in an accessible manner and then establish a set of design requirements for human-centered GNN explanations. Finally, we offer two example prototypes to demonstrate some of those proposed requirements.

Today, cloud workloads are essentially opaque to the cloud platform. Typically, the only information the platform receives is the virtual machine (VM) type and possibly a decoration to the type (e.g., the VM is evictable). Similarly, workloads receive little to no information from the platform; generally, workloads might receive telemetry from their VMs or exceptional signals (e.g., shortly before a VM is evicted). The narrow interface between workloads and platforms has several drawbacks: (1) a surge in VM types and decorations in public cloud platforms complicates customer selection; (2) essential workload characteristics (e.g., low availability requirements, high latency tolerance) are often unspecified, hindering platform customization for optimized resource usage and cost savings; and (3) workloads may be unaware of potential optimizations or lack sufficient time to react to platform events. In this paper, we propose a framework, called Workload Intelligence (WI), for dynamic bi-directional communication between cloud workloads and cloud platform. Via WI, workloads can programmatically adjust their key characteristics, requirements, and even dynamically adapt behaviors like VM priorities. In the other direction, WI allows the platform to programmatically inform workloads about upcoming events, opportunities for optimization, among other scenarios. Because of WI, the cloud platform can drastically simplify its offerings, reduce its costs without fear of violating any workload requirements, and reduce prices to its customers on average by 48.8%.

We propose a coherent transceiver architecture able to transmit information and enhance the security of the optical network by identifying other optical systems and subsystems. Simulations show that identification is obtained with sufficient reliability in standard operating conditions., Comment: The manuscript has been submitted to the 2024 Photonics in Switching and Computing conference

This paper analyses LightGCN in the context of graph recommendation algorithms. Despite the initial design of Graph Convolutional Networks for graph classification, the non-linear operations are not always essential. LightGCN enables linear propagation of embeddings, enhancing performance. We reproduce the original findings, assess LightGCN's robustness on diverse datasets and metrics, and explore Graph Diffusion as an augmentation of signal propagation in LightGCN., Comment: Accepted at NeurIPS'23 Workshop on New in ML; 3 pages, 2 figures, 3 tables

We present a new method for calculating the temperature profile in high explosive (HE) material using a Convolutional Neural Network (CNN). To train/test the CNN, we have developed a hybrid experiment/simulation method for collecting acoustic and temperature data. We experimentally heat cylindrical containers of HE material until detonation/deflagration, where we continuously measure the acoustic bursts through the HE using multiple acoustic transducers lined around the exterior container circumference. However, measuring the temperature profile in the HE in experiment would require inserting a high number of thermal probes, which would disrupt the heating process. Thus, we use two thermal probes, one at the HE center and one at the wall. We then use finite element simulation of the heating process to calculate the temperature distribution, and correct the simulated temperatures based on the experimental center and wall temperatures. We calculate temperature errors on the order of 15{\deg}C, which is approximately 12% of the range of temperatures in the experiment. We also investigate how the algorithm accuracy is affected by the number of acoustic receivers used to collect each measurement and the resolution of the temperature prediction. This work provides a means of assessing the safety status of HE material, which cannot be achieved using existing temperature measurement methods. Additionally, it has implications for range of other applications where internal temperature profile measurements would provide critical information. These applications include detecting chemical reactions, observing thermodynamic processes like combustion, monitoring metal or plastic casting, determining the energy density in thermal storage capsules, and identifying abnormal battery operation.

Real-time video streaming relies on rate control mechanisms to adapt video bitrate to network capacity while maintaining high utilization and low delay. However, the current video rate controllers, such as Google Congestion Control (GCC), are very slow to respond to network changes, leading to link under-utilization and latency spikes. While recent delay-based congestion control algorithms promise high efficiency and rapid adaptation to variable conditions, low-latency video applications have been unable to adopt these schemes due to the intertwined relationship between video encoders and rate control in current systems. This paper introduces Vidaptive, a new rate control mechanism designed for low-latency video applications. Vidaptive decouples packet transmission decisions from encoder output, injecting ``dummy'' padding traffic as needed to treat video streams akin to backlogged flows controlled by a delay-based congestion controller. Vidaptive then adapts the target bitrate of the encoder based on delay measurements to align the video bitrate with the congestion controller's sending rate. Our evaluations atop Google's implementation of WebRTC show that, across a set of cellular traces, Vidaptive achieves ~1.5x higher video bitrate and 1.4 dB higher SSIM, 1.3 dB higher PSNR, and 40% higher VMAF, and it reduces 95th-percentile frame latency by 2.2 s with a slight 17 ms increase in median frame latency.