Atomistic machine learning datasets are increasingly used for training: large immutable snapshots are read repeatedly, shuffled across epochs, staged across clusters' storage systems, and republished as reusable scientific artifacts. This workload differs from interactive scientific curation, where mutable records and ad hoc inspection are often more important than random indexed throughput.
arXiv:2606. 29975v1 Announce Type: new Abstract: Atomistic machine learning datasets are increasingly used for training: large immutable snapshots are read repeatedly, shuffled across epochs, staged across clusters' storage systems, and republished as reusable scientific artifacts.
arXiv:2506. 01883v3 Announce Type: replace-cross Abstract: Training deep learning models on single-cell datasets with hundreds of millions of cells requires loading data from disk, as these datasets exceed available memory.
arXiv:2607. 18187v1 Announce Type: cross Abstract: Large-scale scientific simulations generate volumetric data at rates that far outpace advances in storage and network bandwidth, making effective lossy compression increasingly critical.
The paper introduces NIO Bench, a benchmarking framework that profiles storage I/O for six machine learning model types, using Python hooks and Linux strace to capture detailed access patterns. Experiments on a Ceph-backed Kubernetes cluster show that I/O is dominated by data preparation, model loading, and checkpointing, with training becoming compute-bound once data is staged. The study finds a power‑law distribution of file usage and identifies cache‑miss read tail latency as the main storage bottleneck, recommending aggressive prefetching, page‑cache pinning, and bursty write handling for ML‑optimized storage.
arXiv:2607. 28880v1 Announce Type: cross Abstract: Modern multimedia machine learning workloads increasingly store large-scale datasets in cloud object storage services such as AWS S3.
arXiv:2606. 04557v1 Announce Type: cross Abstract: Large Language Models can reason over long contexts, yet prefilling millions of tokens is wasteful as much of the content remains static across queries.
DeepSeek‑V4.1‑Flash is a multimodal Mixture‑of‑Experts model with 552 B backbone parameters that supports contexts of up to one million tokens. It uses a Causal Encoder‑Decoder architecture that activates 16 B parameters per token during decoding but only 8 B during prefill, improving cost efficiency for agentic workloads. The model introduces cross‑layer KV cache reuse via Compressed Sparse Attention 2 and FP4 KV caching, reducing its global KV cache footprint to 890 bytes per token and its persistent footprint to roughly one‑eighth of the previous version, while still delivering superior performance after pretraining on a 45 T‑token multimodal corpus.
arXiv:2608. 04942v1 Announce Type: cross Abstract: CheMLFlow is an open-source platform for building and executing end-to-end, high-throughput, and agentic workflows for scientific and technological applications.
arXiv:2604.09107v2 Announce Type: replace-cross Abstract: Modern LLM reinforcement learning (RL) workloads require a high-performance weight transfer system to scale training across heterogeneous com...
AtomWorld-Mem is a memory‑restored atomistic world model that reconstructs hidden world states from incomplete crystal snapshots, enabling more accurate long‑horizon atomistic evolution. It uses spatial encoders to capture multi‑scale keyframes and integrates short‑term event memory with long‑term structural memory to predict future states. The restored state guides vacancy‑mediated events in kinetic Monte Carlo simulations, improving progress under fixed event budgets while preserving fidelity across energetic, structural, and transport observables, and it transfers zero‑shot across unseen alloy‑temperature scenarios.
arXiv:2609.39090v1 Announce Type: cross Abstract: Data-driven machine learning (ML) techniques have become an essential tool in many domains of science. Their application to atomistic simulations of...
arXiv:2607. 27273v1 Announce Type: new Abstract: Post-training of large language models is expensive, and existing efficiency improvements mainly focus on selecting informative samples or designing training schedules.