Towards a machine-readable literature: finding relevant papers based on an uploaded powder diffraction pattern

Özer, Berrak; Karlsen, Martin A.; Thatcher, Zachary; Ling, Lan; McMahon, Brian; Strickland, P.R.; Westrip, Simon P.; S., Sang, Koh; Billing, David G.; Ravnsbæk, Dorthe Bomholdt; Billinge, Simon J. L.

doi:10.1107/s2053273322007483

Cited by 3 publications

(2 citation statements)

References 35 publications

(31 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To give more perspective, modern X-ray detectors can generate as many as 1,000,000 images per day, and data post-processing, analysis, and interpretation of the experiments can take over a year [Doucet et al 2020;Wang et al 2017b]. The ability of machine learning methods to process big data, and find patterns in complex data, and the computer vision algorithms for the autonomous detection of images can play a significant role in accelerating the existing workflows at beamline user facilities by providing immediate feedback during the experiments [Wang et al 2017b;Doucet et al 2020;Sullivan et al 2019;Yanxon et al 2023;Wang et al 2017b;Banko et al 2021;Özer et al 2022;Venderley et al 2022].…”

Section: Materials Characterizationmentioning

confidence: 99%

Artificial intelligence in recommender systems

Zhang

Jin

2020

Complex Intell. Syst.

209

View full text Add to dashboard Cite

Recommender systems provide personalized service support to users by learning their previous behaviors and predicting their current preferences for particular products. Artificial intelligence (AI), particularly computational intelligence and machine learning methods and algorithms, has been naturally applied in the development of recommender systems to improve prediction accuracy and solve data sparsity and cold start problems. This position paper systematically discusses the basic methodologies and prevailing techniques in recommender systems and how AI can effectively improve the technological development and application of recommender systems. The paper not only reviews cutting-edge theoretical and practical contributions, but also identifies current research issues and indicates new research directions. It carefully surveys various issues related to recommender systems that use AI, and also reviews the improvements made to these systems through the use of such AI approaches as fuzzy techniques, transfer learning, genetic algorithms, evolutionary algorithms, neural networks and deep learning, and active learning. The observations in this paper will directly support researchers and professionals to better understand current developments and new directions in the field of recommender systems using AI.

show abstract

Section: Materials Characterizationmentioning

confidence: 99%

Artificial intelligence in recommender systems

Zhang

Jin

2020

Complex Intell. Syst.

209

View full text Add to dashboard Cite

show abstract

“…A CSP protocol uses similarity indices to identify duplicate structures, 25 but they are also useful in other contexts such as classification and database searching. [26][27][28][29][30] Similiarity indices based on the comparison of powder diffraction patterns (in the following, "powder-based indices") are attractive because they can be used for both duplicate identification and to compare the candidate structures with experimental XRPD patterns. A further advantage of powder-based indices is that they identify structurally related compounds (conformational phases, isomorphous systems) as similar.…”

Section: Introductionmentioning

confidence: 99%

Powder-Diffraction-Based Structural Comparison for Crystal Structure Prediction without Prior Indexing

Otero de la Roza

2024

Preprint

View full text Add to dashboard Cite

The objective of crystal structure prediction (CSP) is to predict computationally the thermodynamically stable crystal structure of a compound from its stoichiometry or its molecular diagram. Crystal similarity indices measure the degree of similarity between two crystal structures, and are essential in CSP because they are used to identify duplicates. In addition, powder-based indices, which are based on comparing X-ray diffraction patterns, also allow the use of experimental X-ray powder diffraction data to inform the CSP search. Powder-assisted CSP presents two unique difficulties: i) the experimental and computational structures are not entirely comparable because the former is subject to thermal expansion from lattice vibrations, and ii) experimental patterns present features (noise, background contribution, varying peak shapes, etc.) that are not easily predictable computationally. In this work, we present a powder-based similarity index (GPWDF) based on a modification of de Gelder et al.'s index using cross-correlation functions that can be calculated analytically. Based on GPWDF, we also propose a variable-cell similarity index (VC-GPWDF) that assigns a high similarity score to structures that differ only by a lattice deformation and takes advantage of the analytical derivatives of GPWDF with respect to the lattice parameters. VC-GPWDF can be used to identify similarity between: two computational structures generated using different methods, a computational and a experimental structure, and two experimental structures measured under different conditions (e.g. different temperature and pressure). In addition, VC-GPWDF can also be used to compare crystal structures with experimental patterns in combination with an automatic pre-processing step. The proposed similarity indices are simple, efficient, and fully automatic. They require no indexing of the experimental pattern or a guess of the space group, account for deformations caused by varying experimental conditions, give meaningful results even when the experimental pattern is of very poor quality, and have a cost does not increase with the flexibility of the molecular motif.

show abstract

The interoperability of crystallographic data and databases

Brink,

Bruno,

Helliwell

et al. 2024

Int Union Crystallogr J

Self Cite

View full text Add to dashboard Cite

Interoperability of crystallographic data with other disciplines is essential for the smooth and rapid progress of structure-based science in the computer age. Within crystallography and closely related subject areas, there is already a high level of conformance to the generally accepted FAIR principles (that data be findable, accessible, interoperable and reusable) through the adoption of common information exchange protocols by databases, publishers, instrument vendors, experimental facilities and software authors. Driven by the success within these domains, the IUCr has worked closely with CODATA (the Committee on Data of the International Science Council) to help develop the latter's commitment to cross-domain integration of discipline-specific data. The IUCr has, in particular, emphasized the need for standards relating to data quality and completeness as an adjunct to the FAIR data landscape. This can ensure definitive reusable data, which in turn can aid interoperability across domains. A microsymposium at the IUCr 2023 Congress provided an up-to-date survey of data interoperability within and outside of crystallography, expounded using a broad range of examples.

show abstract

Towards a machine-readable literature: finding relevant papers based on an uploaded powder diffraction pattern

Cited by 3 publications

References 35 publications

Artificial intelligence in recommender systems

Artificial intelligence in recommender systems

Powder-Diffraction-Based Structural Comparison for Crystal Structure Prediction without Prior Indexing

The interoperability of crystallographic data and databases

Contact Info

Product

Resources

About