Tootfinder

Context Steering: A New Paradigm for Compression-based Embeddings by Synthesizing Relevant Information Features
Guillermo Sarasa Dur\'an, Ana Granados Fontecha, Francisco de Borja Rodr\'iguez Ort\'iz
https://arxiv.org/abs/2508.14780

Context Steering: A New Paradigm for Compression-based Embeddings by Synthesizing Relevant Information Features
Compression-based distances (CD) offer a flexible and domain-agnostic means of measuring similarity by identifying implicit information through redundancies between data objects. However, as similarity features are derived from the data, rather than defined as an input, it often proves difficult to align with the task at hand, particularly in complex clustering or classification settings. To address this issue, we introduce "context steering," a novel methodology that actively guides the featur…

Boosting Scientific Error-Bounded Lossy Compression through Optimized Synergistic Lossy-Lossless Orchestration
Shixun Wu, Jinwen Pan, Jinyang Liu, Jiannan Tian, Ziwei Qiu, Jiajun Huang, Kai Zhao, Xin Liang, Sheng Di, Zizhong Chen, Franck Cappello
https://arxiv.org/abs/2507.11165

Boosting Scientific Error-Bounded Lossy Compression through Optimized Synergistic Lossy-Lossless Orchestration
As high-performance computing architectures evolve, more scientific computing workflows are being deployed on advanced computing platforms such as GPUs. These workflows can produce raw data at extremely high throughputs, requiring urgent high-ratio and low-latency error-bounded data compression solutions. In this paper, we propose cuSZ-Hi, an optimized high-ratio GPU-based scientific error-bounded lossy compressor with a flexible, domain-irrelevant, and fully open-source framework design. Our n…

Flexible Mixed Precision Quantization for Learned Image Compression
Md Adnan Faisal Hossain, Zhihao Duan, Fengqing Zhu
https://arxiv.org/abs/2506.01221 htt…

Flexible Mixed Precision Quantization for Learned Image Compression
Despite its improvements in coding performance compared to traditional codecs, Learned Image Compression (LIC) suffers from large computational costs for storage and deployment. Model quantization offers an effective solution to reduce the computational complexity of LIC models. However, most existing works perform fixed-precision quantization which suffers from sub-optimal utilization of resources due to the varying sensitivity to quantization of different layers of a neural network. In this p…

Laser-driven bunch compression for ultrashort free-electron laser pulses
Ph. Amstutz, W. Helml, S. Khan, C. Mai, Ch. Gerth, Ch. Mahnke, E. A. Schneidmiller
https://arxiv.org/abs/2508.14592

Laser-driven bunch compression for ultrashort free-electron laser pulses
Generation of ultrashort X-ray pulses in a free-electron laser relies on high-density electron bunches with a precisely adjusted current and energy distribution. To this end, robust and flexible electron bunch manipulation techniques are required that allow a high degree of control over the phase space density of the bunch. This paper reports on the demonstration of ultrashort current spikes with femtosecond duration, created by compressing an electron bunch after a laser-induced energy modulat…

Decomposing Words for Enhanced Compression: Exploring the Number of Runs in the Extended Burrows-Wheeler Transform
Florian Ingels, Ana\"is Denis, Bastien Cazaux
https://arxiv.org/abs/2506.04926

Decomposing Words for Enhanced Compression: Exploring the Number of Runs in the Extended Burrows-Wheeler Transform
The Burrows-Wheeler Transform (BWT) is a fundamental component in many data structures for text indexing and compression, widely used in areas such as bioinformatics and information retrieval. The extended BWT (eBWT) generalizes the classical BWT to multisets of strings, providing a flexible framework that captures many BWT-like constructions. Several known variants of the BWT can be viewed as instances of the eBWT applied to specific decompositions of a word. A central property of the BWT, ess…

PyLate: Flexible Training and Retrieval for Late Interaction Models
Antoine Chaffin, Rapha\"el Sourty
https://arxiv.org/abs/2508.03555 https://arxiv.o…

PyLate: Flexible Training and Retrieval for Late Interaction Models
Neural ranking has become a cornerstone of modern information retrieval. While single vector search remains the dominant paradigm, it suffers from the shortcoming of compressing all the information into a single vector. This compression leads to notable performance degradation in out-of-domain, long-context, and reasoning-intensive retrieval tasks. Multi-vector approaches pioneered by ColBERT aim to address these limitations by preserving individual token embeddings and computing similarity via…