arXiv subjects

Elsa Olivetti

Publications and source records attributed to Elsa Olivetti.

At least 19 recordsLinked to original sources

arXiv2026

Probing Materials Knowledge in LLMs: From Latent Embeddings to Reliable Predictions

Large language models are increasingly applied to materials science, yet fundamental questions remain about their reliability and knowledge encoding. Evaluating 25 LLMs across four materials science tasks -- over 200 base and fine-tuned configurations -- we find that output modality fundamentally determines model behavior. For symbolic tasks, fine-tuning converges to consistent, verifiable answers with reduced response entropy, while for numerical tasks, fine-tuning improves prediction accuracy but models remain inconsistent across repeated inference runs, limiting their reliability as quantitative predictors. For numerical regression, we find that better performance can be obtained by extracting embeddings directly from intermediate transformer layers than from model text output, revealing an ``LLM head bottleneck,'' though this effect is property- and dataset-dependent. Finally, we present a longitudinal study of GPT model performance in materials science, tracking four models over 18 months and observing 9--43\% performance variation that poses reproducibility challenges for scientific applications.

cond-mat.mtrl-sci↗

arXiv2026

PackFlow: Generative Molecular Crystal Structure Prediction via Reinforcement Learning Alignment

Organic molecular crystals underpin technologies ranging from pharmaceuticals to organic electronics, yet predicting solid-state packing of molecules remains challenging because candidate generation is combinatorial and stability is only resolved after costly energy evaluations. Here we introduce PackFlow, a flow matching framework for molecular crystal structure prediction (CSP) that generates heavy-atom crystal proposals by jointly sampling Cartesian coordinates and unit-cell lattice parameters given a molecular graph. This lattice-aware generation interfaces directly with downstream relaxation and lattice-energy ranking, positioning PackFlow as a scalable proposal engine within standard CSP pipelines. To explicitly steer generation toward physically favourable regions, we propose physics alignment, a reinforcement learning post-training stage that uses machine-learned interatomic potential energies and forces as stability proxies. Physics alignment improves physical validity without altering inference-time sampling. We validate PackFlow's performance against heuristic baselines through two distinct evaluations. First, on a broad unseen set of molecular systems, we demonstrate superior candidate generation capability, with proposals exhibiting greater structural similarity to experimental polymorphs. Second, we assess the full end-to-end workflow on two unseen CSP blind-test case studies, including relaxation and lattice-energy analysis. In both settings, PackFlow outperforms heuristics-based methods by concentrating probability mass in low-energy basins, yielding candidates that relax into lower-energy minima and offering a practical route to amortize the relax-and-rank bottleneck.

physics.chem-ph↗

arXiv2025

Masked Mineral Modeling: Continent-Scale Mineral Prospecting via Geospatial Infilling

Minerals play a critical role in the advanced energy technologies necessary for decarbonization, but characterizing mineral deposits hidden underground remains costly and challenging. Inspired by recent progress in generative modeling, we develop a learning method which infers the locations of minerals by masking and infilling geospatial maps of resource availability. We demonstrate this technique using mineral data for the conterminous United States, and train performant models, with the best achieving Dice coefficients of $0.31 \pm 0.01$ and recalls of $0.22 \pm 0.02$ on test data at 1$\times$1 mi$^2$ spatial resolution. One major advantage of our approach is that it can easily incorporate auxiliary data sources for prediction which may be more abundant than mineral data. We highlight the capabilities of our model by adding input layers derived from geophysical sources, along with a nation-wide ground survey of soils originally intended for agronomic purposes. We find that employing such auxiliary features can improve inference performance, while also enabling model evaluation in regions with no recorded minerals.

stat.ML↗

arXiv2025

DiffSyn: A Generative Diffusion Approach to Materials Synthesis Planning

The synthesis of crystalline materials, such as zeolites, remains a significant challenge due to a high-dimensional synthesis space, intricate structure-synthesis relationships and time-consuming experiments. Considering the one-to-many relationship between structure and synthesis, we propose DiffSyn, a generative diffusion model trained on over 23,000 synthesis recipes spanning 50 years of literature. DiffSyn generates probable synthesis routes conditioned on a desired zeolite structure and an organic template. DiffSyn achieves state-of-the-art performance by capturing the multi-modal nature of structure-synthesis relationships. We apply DiffSyn to differentiate among competing phases and generate optimal synthesis routes. As a proof of concept, we synthesize a UFI material using DiffSyn-generated synthesis routes. These routes, rationalized by density functional theory binding energies, resulted in the successful synthesis of a UFI material with a high Si/Al$_{\text{ICP}}$ of 19.0, which is expected to improve thermal stability and is higher than that of any previously recorded.

cond-mat.mtrl-sci↗

arXiv2025

The Future of Artificial Intelligence and the Mathematical and Physical Sciences (AI+MPS)

This community paper developed out of the NSF Workshop on the Future of Artificial Intelligence (AI) and the Mathematical and Physics Sciences (MPS), which was held in March 2025 with the goal of understanding how the MPS domains (Astronomy, Chemistry, Materials Research, Mathematical Sciences, and Physics) can best capitalize on, and contribute to, the future of AI. We present here a summary and snapshot of the MPS community's perspective, as of Spring/Summer 2025, in a rapidly developing field. The link between AI and MPS is becoming increasingly inextricable; now is a crucial moment to strengthen the link between AI and Science by pursuing a strategy that proactively and thoughtfully leverages the potential of AI for scientific discovery and optimizes opportunities to impact the development of AI by applying concepts from fundamental science. To achieve this, we propose activities and strategic priorities that: (1) enable AI+MPS research in both directions; (2) build up an interdisciplinary community of AI+MPS researchers; and (3) foster education and workforce development in AI for MPS researchers and students. We conclude with a summary of suggested priorities for funding agencies, educational institutions, and individual researchers to help position the MPS community to be a leader in, and take full advantage of, the transformative potential of AI+MPS.

cs.AI↗

arXiv2025

The Cloud Next Door: Investigating the Environmental and Socioeconomic Strain of Datacenters on Local Communities

Datacenters have become the backbone of modern digital infrastructure, powering the rapid rise of artificial intelligence and promising economic growth and technological progress. However, this expansion has brought growing tensions in the local communities where datacenters are already situated or being proposed. While the mainstream discourse often focuses on energy usage and carbon footprint of the computing sector at a global scale, the local socio-environmental consequences -- such as health impacts, water usage, noise pollution, infrastructural strain, and economic burden -- remain largely underexplored and poorly addressed. In this work, we surface these community-level consequences through a mixed-methods study that combines quantitative data with qualitative insights. Focusing on Northern Virginia's ``Data Center Valley,'' we highlight how datacenter growth reshapes local environments and everyday life, and examine the power dynamics that determine who benefits and who bears the costs. Our goal is to bring visibility to these impacts and prompt more equitable and informed decisions about the future of digital infrastructure.

cs.DC↗

arXiv2025

Contrastive Learning of English Language and Crystal Graphs for Multimodal Representation of Materials Knowledge

Artificial intelligence (AI) is increasingly used for the inverse design of materials, such as crystals and molecules. Existing AI research on molecules has integrated chemical structures of molecules with textual knowledge to adapt to complex instructions. However, this approach has been unattainable for crystals due to data scarcity from the biased distribution of investigated crystals and the lack of semantic supervision in peer-reviewed literature. In this work, we introduce a contrastive language-crystals model (CLaC) pre-trained on a newly synthesized dataset of 126k crystal structure-text pairs. To demonstrate the advantage of using synthetic data to overcome data scarcity, we constructed a comparable dataset extracted from academic papers. We evaluate CLaC's generalization ability through various zero-shot cross-modal tasks and downstream applications. In experiments, CLaC achieves state-of-the-art zero-shot generalization performance in understanding crystal structures, surpassing latest large language models.

cs.CL↗

arXiv2025

Retro-Rank-In: A Ranking-Based Approach for Inorganic Materials Synthesis Planning

Retrosynthesis strategically plans the synthesis of a chemical target compound from simpler, readily available precursor compounds. This process is critical for synthesizing novel inorganic materials, yet traditional methods in inorganic chemistry continue to rely on trial-and-error experimentation. Emerging machine-learning approaches struggle to generalize to entirely new reactions due to their reliance on known precursors, as they frame retrosynthesis as a multi-label classification task. To address these limitations, we propose Retro-Rank-In, a novel framework that reformulates the retrosynthesis problem by embedding target and precursor materials into a shared latent space and learning a pairwise ranker on a bipartite graph of inorganic compounds. We evaluate Retro-Rank-In's generalizability on challenging retrosynthesis dataset splits designed to mitigate data duplicates and overlaps. For instance, for Cr2AlB2, it correctly predicts the verified precursor pair CrB + Al despite never seeing them in training, a capability absent in prior work. Extensive experiments show that Retro-Rank-In sets a new state-of-the-art, particularly in out-of-distribution generalization and candidate set ranking, offering a powerful tool for accelerating inorganic material synthesis.

physics.chem-ph↗

arXiv2024

The Sunk Carbon Fallacy: Rethinking Carbon Footprint Metrics for Effective Carbon-Aware Scheduling

The rapid increase in computing demand and its corresponding energy consumption have focused attention on computing's impact on the climate and sustainability. Prior work proposes metrics that quantify computing's carbon footprint across several lifecycle phases, including its supply chain, operation, and end-of-life. Industry uses these metrics to optimize the carbon footprint of manufacturing hardware and running computing applications. Unfortunately, prior work on optimizing datacenters' carbon footprint often succumbs to the \emph{sunk cost fallacy} by considering embodied carbon emissions (a sunk cost) when making operational decisions (i.e., job scheduling and placement), which leads to operational decisions that do not always reduce the total carbon footprint. In this paper, we evaluate carbon-aware job scheduling and placement on a given set of servers for a number of carbon accounting metrics. Our analysis reveals state-of-the-art carbon accounting metrics that include embodied carbon emissions when making operational decisions can actually increase the total carbon footprint of executing a set of jobs. We study the factors that affect the added carbon cost of such suboptimal decision-making. We then use a real-world case study from a datacenter to demonstrate how the sunk carbon fallacy manifests itself in practice. Finally, we discuss the implications of our findings in better guiding effective carbon-aware scheduling in on-premise and cloud datacenters.

cs.DC↗

arXiv2024

Scoping Sustainable Collaborative Mixed Reality

Mixed Reality (MR) is becoming ubiquitous as it finds its applications in education, healthcare, and other sectors beyond leisure. While MR end devices, such as headsets, have low energy intensity, the total number of devices and resource requirements of the entire MR ecosystem, which includes cloud and edge endpoints, can be significant. The resulting operational and embodied carbon footprint of MR has led to concerns about its environmental implications. Recent research has explored reducing the carbon footprint of MR devices by exploring hardware design space or network optimizations. However, many additional avenues for enhancing MR's sustainability remain open, including energy savings in non-processor components and carbon-aware optimizations in collaborative MR ecosystems. In this paper, we aim to identify key challenges, existing solutions, and promising research directions for improving MR sustainability. We explore adjacent fields of embedded and mobile computing systems for insights and outline MR-specific problems requiring new solutions. We identify the challenges that must be tackled to enable researchers, developers, and users to avail themselves of these opportunities in collaborative MR systems.

cs.CY↗

arXiv2024

Accelerating Discovery of Metal-Insulator Transition Compounds Using Physics-Informed Machine Learning

Metal-insulator transition (MIT) materials are a useful platform for emerging microelectronic, optoelectronic, and neuromorphic devices, but their discovery is hindered by the high computational cost of electronic structure modeling, the complexity of underlying mechanisms, and the challenges of experimental validation. Here, we present a physics-informed machine learning framework that accelerates the discovery of thermally driven MIT materials. Using a trained classifier, we screen a crystal structure database to identify promising candidates for higher fidelity simulations. We focus on Ca$_2$Fe$_3$O$_8$, CaCo$_2$O$_4$, and CaMn$_2$O$_4$, and use density functional theory (DFT) to determine their electronic and magnetic ground states and assess their microscopic MIT mechanisms. We further apply machine learning regression models to estimate their transition temperatures and employ synthesis prediction tools to identify likely precursors and reaction routes. This integrated approach reduces the time and effort required to identify, understand, and synthesize new MIT materials, providing a generalizable pathway for accelerating correlated quantum materials discovery.

cond-mat.mtrl-sci↗

arXiv2022

MatKG: The Largest Knowledge Graph in Materials Science -- Entities, Relations, and Link Prediction through Graph Representation Learning

This paper introduces MatKG, a novel graph database of key concepts in material science spanning the traditional material-structure-property-processing paradigm. MatKG is autonomously generated through transformer-based, large language models and generates pseudo ontological schema through statistical co-occurrence mapping. At present, MatKG contains over 2 million unique relationship triples derived from 80,000 entities. This allows the curated analysis, querying, and visualization of materials knowledge at unique resolution and scale. Further, Knowledge Graph Embedding models are used to learn embedding representations of nodes in the graph which are used for downstream tasks such as link prediction and entity disambiguation. MatKG allows the rapid dissemination and assimilation of data when used as a knowledge base, while enabling the discovery of new relations when trained as an embedding model.

cond-mat.mtrl-sci↗

arXiv2022

Deep Reinforcement Learning for Inverse Inorganic Materials Design

A major obstacle to the realization of novel inorganic materials with desirable properties is the inability to perform efficient optimization across both materials properties and synthesis of those materials. In this work, we propose a reinforcement learning (RL) approach to inverse inorganic materials design, which can identify promising compounds with specified properties and synthesizability constraints. Our model learns chemical guidelines such as charge and electronegativity neutrality while maintaining chemical diversity and uniqueness. We demonstrate a multi-objective RL approach, which can generate novel compounds with targeted materials properties including formation energy and bulk/shear modulus alongside a lower sintering temperature synthesis objectives. Using this approach, the model can predict promising compounds of interest, while suggesting an optimized chemical design space for inorganic materials discovery.

cond-mat.mtrl-sci↗

arXiv2022

Data-driven prediction of room temperature density for multicomponent silicate-based glasses

Density is one of the most commonly measured or estimated materials properties, especially for glasses and melts that are of significant interest to many fields, including metallurgy, geology, materials science and sustainable cements. Here, two types of machine learning (ML) models (i.e., random forest (RF) and artificial neural network (ANN)) have been developed to predict the room-temperature density of glasses in the compositional space of CaO-MgO-Al2O3-SiO2-TiO2-FeO-Fe2O3-Na2O-K2O-MnO (CMASTFNKM), based on ~2100 data points mined from ~140 literature studies. The results show that the RF and ANN models give accurate predictions of glass density with R2 values, RMSE, and MAPE of ~0.96-0.98, ~0.02-0.03 g/cm3 and ~0.59-0.79%, respectively, for the 15% testing set, which are more accurate compared with empirical density models based on ionic packing ratio (with R2 values, RMSE, and MAPE of ~0.28-0.91, ~0.05-0.15 g/cm3, and ~1.40-4.61%, respectively). Furthermore, glass density is shown to be a reliable reactivity indicator for a range of CaO-Al2O3-SiO2 (CAS) and volcanic glasses due to its strong correlation (R2 values above ~0.90) with the average metal-oxygen dissociation energy (a structural descriptor) of these glasses. Analysis of the predicted density-composition relationships from these models (for selected compositional subspaces) suggests that the ANN model exhibits a certain level of transferability (i.e., ability to extrapolate to compositional space not (or less) covered in the database) and captures known features including the mixed alkaline earth effects for (CaO-MgO)0.5-(Al2O3-SiO2)0.5 glasses.

cond-mat.mtrl-sci↗

arXiv2022

Augmenting Scientific Creativity with Retrieval across Knowledge Domains

Exposure to ideas in domains outside a scientist's own may benefit her in reformulating existing research problems in novel ways and discovering new application domains for existing solution ideas. While improved performance in scholarly search engines can help scientists efficiently identify relevant advances in domains they may already be familiar with, it may fall short of helping them explore diverse ideas \textit{outside} such domains. In this paper we explore the design of systems aimed at augmenting the end-user ability in cross-domain exploration with flexible query specification. To this end, we develop an exploratory search system in which end-users can select a portion of text core to their interest from a paper abstract and retrieve papers that have a high similarity to the user-selected core aspect but differ in terms of domains. Furthermore, end-users can `zoom in' to specific domain clusters to retrieve more papers from them and understand nuanced differences within the clusters. Our case studies with scientists uncover opportunities and design implications for systems aimed at facilitating cross-domain exploration and inspiration.

cs.IR↗

arXiv2021

Inorganic Synthesis Reaction Condition Prediction with Generative Machine Learning

Data-driven synthesis planning with machine learning is a key step in the design and discovery of novel inorganic compounds with desirable properties. Inorganic materials synthesis is often guided by chemists' prior knowledge and experience, built upon experimental trial-and-error that is both time and resource consuming. Recent developments in natural language processing (NLP) have enabled large-scale text mining of scientific literature, providing open source databases of synthesis information of synthesized compounds, material precursors, and reaction conditions (temperatures, times). In this work, we employ a conditional variational autoencoder (CVAE) to predict suitable inorganic reaction conditions for the crucial inorganic synthesis steps of calcination and sintering. We find that the CVAE model is capable of learning subtle differences in target material composition, precursor compound identities, and choice of synthesis route (solid-state, sol-gel) that are present in the inorganic synthesis space. Moreover, the CVAE can generalize well to unseen chemical entities and shows promise for predicting reaction conditions for previously unsynthesized compounds of interest.

cond-mat.mtrl-sci↗

arXiv2021

Development of structural descriptors to predict dissolution rate of volcanic glasses: molecular dynamic simulations

Establishing the composition-structure-property relationships for amorphous materials is critical for many important natural and engineering processes, including the dissolution of highly complex volcanic glasses. In this investigation, we performed force field molecular dynamics (MD) simulations to generate detailed structural representations for ten natural CaO-MgO-Al2O3-SiO2-TiO2-FeO-Fe2O3-Na2O-K2O glasses with compositions ranging from rhyolitic to basaltic. Based on the resulting atomic structural representations at 300 K, we have calculated the partial radial distribution functions, nearest interatomic distances and coordination number, which are consistent with the literature data on silicate-based glasses. Based on these structural attributes and classical bond valence models, we have introduced a novel structural descriptor, i.e., average metal-oxygen (M-O) bond strength parameter, which has captured the log dissolution rates of the ten glasses at both acidic and basic conditions (based on literature data) with R2 values of ~0.80-0.92 based on linear regression. This structural descriptor is seen to outperform several other structural descriptors also derived from MD simulation results, including the average metal oxide dissociation energy, the average self-diffusion coefficient of all the atoms at their melting points, and the energy barrier of self-diffusion. Furthermore, we showed that the MD-derived descriptors generally exhibit better predictive performance than the degree of depolymerization parameter commonly used to describe glass and mineral reactivity. The results suggest that the structural descriptors derived from MD simulations, especially the average M-O bond strength parameter, are promising structural descriptors for connecting composition with dissolution rates of highly complex natural glasses.

cond-mat.mtrl-sci↗

arXiv2019

The Materials Science Procedural Text Corpus: Annotating Materials Synthesis Procedures with Shallow Semantic Structures

Materials science literature contains millions of materials synthesis procedures described in unstructured natural language text. Large-scale analysis of these synthesis procedures would facilitate deeper scientific understanding of materials synthesis and enable automated synthesis planning. Such analysis requires extracting structured representations of synthesis procedures from the raw text as a first step. To facilitate the training and evaluation of synthesis extraction models, we introduce a dataset of 230 synthesis procedures annotated by domain experts with labeled graphs that express the semantics of the synthesis sentences. The nodes in this graph are synthesis operations and their typed arguments, and labeled edges specify relations between the nodes. We describe this new resource in detail and highlight some specific challenges to annotating scientific text with shallow semantic structure. We make the corpus available to the community to promote further research and development of scientific information extraction systems.

cs.CL↗