23 papers · ranked by Valyu relevance
Defu Liu, Yixiao Zhu, Zhe Liu, Yi Liu + 4 more
'Ruihao Li' 'Wei Yi'] The exceptional performance of general-purpose large models has driven various industries to focus on developing domain-specific models. However, large models are not only time-consuming and labor-intensive during the training phase but also have very high hardware requirements during the…
Moshik Hershcovitch, Andrew W. Wood, Leshem Choshen, Guy Girmonsky + 6 more
'Roy Leibovitz' 'Ilias Ennmouri' 'Michal Malka' 'Peter Chin' 'Sundararaman Swaminathan' 'Danny Harnik'] With the growth of model sizes and the scale of their deployment, their sheer size burdens the infrastructure requiring more network and more storage to accommodate these. While there is a vast model compression…
Angie Boggust, Venkatesh Sivaraman, Yannick Assogba, Donghao Ren + 2 more
Across ML Model Compression Experiments Authors: Angie Boggust, Venkatesh Sivaraman, Yannick Assogba, Donghao Ren, Dominik Moritz, Fred Hohman Title: Compress and Compare: Interactively Evaluating Efficiency and Behavior Across ML Model Compression Experiments Authors: Angie Boggust, Venkatesh Sivaraman, Yannick…
Moshik Hershcovitch, Leshem Choshen, Andrew W. Wood, Ilias Enmouri + 3 more
'Peter Chin' 'Sundararaman Swaminathan' 'Danny Harnik'] Abstract—With the growth of model sizes and scale of their deployment, their sheer size burdens the infrastructure requiring more network and more storage to accommodate these. While there is a vast literature about reducing model sizes, we investigate a more…
Samir Brahim Belhaouari, Insaf Kraidia
Large Language Models (LLMs) have revolutionized artificial intelligence by enabling multitasking across diverse fields. However, their high computational demands result in significant environmental impacts, particularly in terms of energy and water consumption. This paper addresses these issues by proposing an…
Fatemeh Kamali, Amir Abolfazl Suratgar, Mohammadbagher Menhaj, Reza Abbasi-Asl
Voxelwise encoding models based on convolutional neural networks (CNNs) have emerged as state-of-the-art predictive models of brain activity evoked by natural movies. Despite their superior predictive performance, the huge number of parameters in CNN-based models have made them difficult to interpret. Here, we…
Omri Armstrong, Ran Gilad-Bachrach
Machine Learning models should ideally be compact and robust. Compactness provides efficiency and comprehensibility whereas robustness provides resilience. Both topics have been studied in recent years but in isolation. Here we present a robust model compression scheme which is independent of model types: it can…
Fatemeh Kamali, Amir Abolfazl Suratgar, Mohammadbagher Menhaj, Reza Abbasi-Asl + 1 more
'Reza Abbasi-Asl' 'Daniele Marinazzo'] Voxelwise encoding models based on convolutional neural networks (CNNs) are widely used as predictive models of brain activity evoked by natural movies. Despite their superior predictive performance, the huge number of parameters in CNN-based models have made them difficult to…
Yuheng Bu, Weihao Gao, Shaofeng Zou, Venugopal V. Veeravalli + 2 more
'Lizhong Zheng' 'Chao Tian'] It has been reported in many recent works on deep model compression that the population risk of a compressed model can be even better than that of the original model. In this paper, an information-theoretic explanation for this population risk improvement phenomenon is provided by jointly…
Marek Kokot, Amitava Roy, Travis J Wheeler, Sebastian Deorowicz
Molecular dynamics (MD) simulations model the physical movements of atoms in biomolecular systems over time, providing atomic-resolution insight into conformational changes, binding events, and dynamic behaviors that cannot be captured by static structures alone. As such, MD simulations are playing an increasingly…
Authors not listed
Efficient lossless compression is essential for minimizing storage costs and transmission overhead while preserving data integrity. Traditional compression techniques, such as dictionary-based and statistical methods, often struggle to optimally exploit the structure and redundancy in complex data formats. Recent…
Jonathan Warrell, Hussein Mohsen, Mark Gerstein
Deep learning methods have achieved state-of-the-art performance in many domains of artificial intelligence, but are typically hard to interpret. Network interpretation is important for multiple reasons, including knowledge discovery, hypothesis generation, fairness and establishing trust. Model transformations provide…
Runzhao Yang, Tingxiong Xiao, Yuxiao Cheng, Anan Li + 7 more
Efficient storage and sharing of massive biomedical data would open up their wide accessibility to different institutions and disciplines. However, compressors tailored for natural photos/videos are rapidly limited for biomedical data, while emerging deep learning based methods demand huge training data and are…
Swathi Shree Narashiman, Nitin Chandrachoodan
Data compression continues to evolve, with traditional information theory methods being widely used for compressing text, images, and videos. Recently, there has been growing interest in leveraging Generative AI for predictive compression techniques. This paper 1 introduces a lossless text compression approach using a…
Felix Zeller, Chieh-Min Hsieh, Wilke Dononelli, Tim Neudecker
The field of liquid-phase and solid-state high-pressure chemistry has exploded since the advent of the diamond anvil cell, an experimental technique that allows the application of pressures up to several hundred gigapascal. To complement high-pressure experiments, a large number of computational tools have been…
Sibusiso B. Buthelezi, Jules R. Tapamo, Nikolaos Mitianoudis
We present a hybrid end-to-end learned image compression framework that combines a CNN-based variational autoencoder (VAE) with an efficient hierarchical Swin Transformer to address the limitations of existing entropy models in capturing global dependencies under computational constraints. Traditional VAE-based codecs…
Gregory P. Way, Michael Zietz, Daniel S. Himmelstein, Casey S. Greene
Unsupervised machine learning algorithms applied to gene expression data extract latent, or hidden, signals representing technical and biological sources of variation. However, these algorithms require a user to select a biologically-appropriate latent dimensionality. We compressed gene expression data from three large…
Anders Andreasen, Maria Bonto, Fernando Montero
– This paper presents a framework for optimisation and techno-economic analysis of various pressurisation pathways for CO2 pipeline transportation. The pressurisation pathways include a conventional compression only case from initial to final pressure, a sub-critical compression part followed by cooling, liquefaction…
Daniel Probst
Last year, a preprint gained notoriety, proposing that a k-nearest neighbour classifier is able to outperform large-language models using compressed text as input and normalised compression distance (NCD) as a metric. In chemistry and biochemistry, molecules are often represented as strings, such as SMILES for small…
Shubham Chandak, Kedar Tatwawadi, Srivatsan Sridhar, Tsachy Weissman
Nanopore sequencing provides a real-time and portable solution to genomic sequencing, with long reads enabling better assembly and structural variant discovery than second generation technologies. The nanopore sequencing process generates huge amounts of data in the form of raw current data, which must be compressed to…
David Podgorelec, Damjan Strnad, Ivana Kolingerová, Borut Žalik + 1 more
'Jun Chen'] After a boom that coincided with the advent of the internet, digital cameras, digital video and audio storage and playback devices, the research on data compression has rested on its laurels for a quarter of a century. Domain-dependent lossy algorithms of the time, such as JPEG, AVC, MP3 and others…
Authors not listed
Cost effective and reliable hydrogen compression remains a challenging barrier in the wide-spread adoption of hydrogen as an energy carrier. The prevailing technology of mechanical compression suffers from several drawbacks, some of which can be addressed by non-mechanical compression strategies (e.g., electrochemical…
Authors not listed
Understanding the complex chemistry of organic materials under dynamic compression is important for many applications, but is challenging due to the large number of reactions occurring at various time scales. Here, we develop a machine learning potential based on Chebyshev polynomials to study the insensitive energetic…