22 papers · ranked by Valyu relevance
Gyung Hyun Je, Colin Raffel
While large language models (LLMs) demonstrate reasonable zero-shot capability across many downstream tasks, fine-tuning is a common practice to improve their performance. However, a task's data efficiency — i.e., the number of fine-tuning examples needed to achieve a desired level of performance — is often unknown…
Lin Cai, Xuzhong Qu, Hang Zhou, Ning Li + 8 more
High-resolution large-volume biological imaging techniques are now widely used in biological research, but inefficiencies in data processing and visualization persist due to bottlenecks in loading/saving pipelines and limitations of conventional pyramid formats. To address these issues, we have decoupled data loading…
Panagiota Nikolaou, Freddy Gabbay, Jawad Haj-Yahya, Yiannakis Sazeides
—This work aims to improve a data center's efficiency by optimizing the server upgrade plan: determine the optimal timing for replacing old servers with new ones. The opportunity presented by this approach is demonstrated through a study based on historical server data. The study establishes a significant opportunity…
Charis Xuan Xie, Alice-Maria Toader, Anna De Simoni, Sandra Eldridge + 3 more
Background Randomised implementation trials evaluate the effects of implementation strategies on implementation outcomes and may also monitor clinical effectiveness. Routine healthcare data are used in implementation trials for participant identification, intervention delivery, and/or outcome ascertainment. Trial…
Ireen J. Kal, Lidwien P. Smabers, Joanna von Berg, Giulia Perticari + 6 more
For the organizing committee of a challenge: 1. Start with thorough inventory of storage entities used by the target audience. During the challenges, this inventory was done by the organizers. This ensures that no relevant locations are overlooked and enables the development of data collection forms that are…
Vanessa Schmidt, Huy Hoang Nguyen, Cédric Jung, Shirin Salehi + 1 more
Resource constraints increasingly determine what can be trained, fine-tuned, and deployed in large language models (LLMs), yet efficiency is often studied through isolated techniques rather than as an interacting system of limits. This survey adopts a constraint-centric perspective and organizes recent progress around…
William Kuszmaul, Jingxun Liang, Renfei Zhou
bits in the regime where U = poly(n). This resolves an open question by Pibiri and Venturini as to whether a redundancy (a.k.a. space overhead) of o(n) bits is possible, and is the first dynamic solution to bypass the so-called tree-structure bottleneck, in which the bits needed to encode some dynamic tree structure…
Oderinwale, Hamidah, Kazlauskas, Anna
Despite data's central role in AI production, it remains the least understood input. As AI labs exhaust public data and turn to proprietary sources with deals reaching hundreds of millions of dollars, research across computer science, economics, law, and policy has fragmented. We establish data economics as a coherent…
Adrian Tkachenko, Sepehr Salem, Ayotomiwa Ezekiel Adeniyi, Zülal Bingöl + 6 more
High-throughput sequencing (HTS) enables population-scale genomics but generates massive datasets, creating bottlenecks in storage, transfer, and analysis. FASTQ, the standard format for over two decades, stores one byte per base and one byte per quality score, leading to inefficient I/O, high storage costs, and…
Anastasia C. Diseth, Simon J. Puglisi
Given a sequence S of subsets of symbols drawn from an alphabet of size σ, a subset rank query srank(i, c) asks for the number of subsets before the ith subset that contain the symbol c. It was recently shown (Alanko et al., Proc. SIAM ACDA, 2023) that subset rank queries on the spectral Burrows-Wheeler lead to…
Vlad C. Coroamă, Oana Dumbravă
This report analyses data centre (DC) sustainability and circularity, revealing existing synergies and trade-offs: The PUE is too coarse, mixing cooling and power provisioning. It wrongly attributes server fan consumption and transformation losses to IT energy. It does not measure compute but infrastructure efficiency…
Niangen Ye, Jiawen Zhu, Baojun Chen, Dong Wang + 3 more
Communication is pivotal in LLM training, and a thorough analysis of the communication efficiency of AI data center (AIDC) network is essential for guiding the design of these capital-intensive clusters. However, conventional metrics are inadequate for such analysis, as they do not directly link network activity to…
Marwa E. Madkour, Salah E. Soliman, Moawad I. Dessouky, Fathi E. Abd El-Samie + 2 more
This research paper presents an efficient data collection scheme for Wireless Sensor Networks (WSNs) that simultaneously compresses and encrypts sensor data to extend network lifespan. To address WSN resource limitations, the scheme combines Compressive Sensing (CS) with Elliptic Curve Cryptography (ECC) and Elliptic…
Richard Durbin
Skiplists (Pugh, 1990) are probabilistic data structures over ordered lists supporting 𝒪 (log N) insertion and search, which share many properties with balanced binary trees. Previously we introduced the graph Burrows-Wheeler transform (GBWT) to support efficient search over pangenome path sets, but current…
Florian Ingels, Léa Vandamme, Mathilde Girard, Clément Agret + 2 more
Modern sequencing continues to drive explosive growth of nucleotide sequence archives, pushing MinHash-derived sketching methods to their practical scalability limits. State-of-the-art tools such as Mash, Dashing2, and Bindash2 provide compact sketches and accurate similarity estimates for large collections, yet…
Enrico Seiler, Myrthe Willemsen, Vitor C. Piro, Knut Reinert
A continued decrease in sequencing costs has facilitated the exponential increase in available sequencing data, with public databases like the European Nucleotide Archive (ENA) and Sequence Read Archive (SRA) reaching well in the order of petabases [1, 2]. This has been the incentive to develop more scalable tools for…
C. B. Lv, Huan Li, Dingyu Yang, Zhongle Xie + 2 more
With streaming floating-point numbers being increasingly prevalent, effective and efficient compression of such data is critical. Compression schemes must be able to exploit the similarity, or smoothness, of consecutive numbers and must be able to contend with extreme conditions, such as high-precision values or the…
N. Mookhambika, J. Raja, K. Srihari
In a sensor network, a sensor node does not aggregate the data packets efficiently, since the packet is dropped for poor connection between the sensor nodes in the network. The sensor node location is varied in every area, so infrequent to share the information with its neighbor by the sender node. The communication is…
Authors not listed
The concept of thermodynamic free energy plays a central role across various disciplines, providing critical insights into system dynamics, energy flow, and sustainability. In materials science, Gibbs free energies are typically determined through phase diagram modeling, which relies on a broad range of thermodynamic…
Rodrigo Torres-Avilés, Gilberto Gutiérrez, Martita Muñoz, Mónica Caniupán + 1 more
This study aims to explore the feasibility of executing Skyline Queries-used to identify optimal data points based on multiple criteria-within compact data structures. While existing algorithms typically operate on plain or indexed datasets, this work investigates how such queries can be efficiently resolved using…
Authors not listed
Biological processes underpin centralized wastewater treatment but are difficult to deploy at small scale. Thermomechanical and thermochemical approaches could enable household-level sanitation, yet their economic and environmental potential remains unclear. We assessed two prototype household reinvented toilets…
Sergey Kozlov, Elena Spirina, Alexey Ulanov, Claudio Savaglio
In this study, a technique for applying the collective dynamic routing method for jointly managing data flows across a heterogeneous network serving both mobile users and sensors simultaneously is developed. A heterogeneous network structure is proposed, including 802.15.4, Wi-Fi 2.4, and Wi-Fi 5 subnets. To implement…