arXiv subjects

Tiago Botari

Publications and source records attributed to Tiago Botari.

13 recordsLinked to original sources

arXiv2026

Lens-descriptor guided evolutionary algorithm for optimization of complex optical systems with glass choice

Designing high-performance optical lenses entails exploring a high-dimensional, tightly constrained space of surface curvatures, glass choices, element thicknesses, and spacings. In practice, standard optimizers (e.g., gradient-based local search and evolutionary strategies) often converge to a single local optimum, overlooking many comparably good alternatives that matter for downstream engineering decisions. We propose the Lens Descriptor-Guided Evolutionary Algorithm (LDG-EA), a two-stage framework for multimodal lens optimization. LDG-EA first partitions the design space into behavior descriptors defined by curvature-sign patterns and material indices, then learns a probabilistic model over descriptors to allocate evaluations toward promising regions. Within each descriptor, LDG-EA applies the Hill-Valley Evolutionary Algorithm with covariance-matrix self-adaptation to recover multiple distinct local minima, optionally followed by gradient-based refinement. On a 24-variable (18 continuous and 6 integer), six-element Double-Gauss topology, LDG-EA generates on average around 14500 candidate minima spanning 636 unique descriptors, an order of magnitude more than a CMA-ES baseline, while keeping wall-clock time at one hour scale. Although the best LDG-EA design is slightly worse than a fine-tuned reference lens, it remains in the same performance range. Overall, the proposed LDG-EA produces a diverse set of solutions while maintaining competitive quality within practical computational budgets and wall-clock time.

cs.NE↗

arXiv2021

Machine learning unveils composition-property relationships in chalcogenide glasses

Due to their unique optical and electronic functionalities, chalcogenide glasses are materials of choice for numerous microelectronic and photonic devices. However, to extend the range of compositions and applications, profound knowledge about composition-property relationships is necessary. To this end, we collected a large quantity of composition-property data on chalcogenide glasses from SciGlass database regarding glass transition temperature ($T_g$), Young's modulus ($E$), coefficient of thermal expansion (CTE), and refractive index ($n_D$). With these data, we induced predictive models using three machine learning algorithms: Random Forest, K-nearest Neighbors, and Classification and Regression Trees. Finally, the induced models were interpreted by computing the SHAP (SHapley Additive exPlanations) values of the chemical features, which revealed the key elements that significantly impacted the tested properties and quantified their impact. For instance, Ge and Ga increase $T_g$ and $E$ and decrease CTE (three properties that depend on bond strength), whereas Se has the opposite effect. Te, As, Tl, and Sb increase $n_D$ (which strongly depends on polarizability), whereas S, Ge, and P diminish it. Knowledge about the effect of each element on the glass properties is precious for semi-empirical compositional development trials or simulation-driven formulations. The induced models can be used to design novel chalcogenide glasses with required combinations of properties.

cond-mat.mtrl-sci↗

arXiv2020

MeLIME: Meaningful Local Explanation for Machine Learning Models

Most state-of-the-art machine learning algorithms induce black-box models, preventing their application in many sensitive domains. Hence, many methodologies for explaining machine learning models have been proposed to address this problem. In this work, we introduce strategies to improve local explanations taking into account the distribution of the data used to train the black-box models. We show that our approach, MeLIME, produces more meaningful explanations compared to other techniques over different ML models, operating on various types of data. MeLIME generalizes the LIME method, allowing more flexible perturbation sampling and the use of different local interpretable models. Additionally, we introduce modifications to standard training algorithms of local interpretable models fostering more robust explanations, even allowing the production of counterfactual examples. To show the strengths of the proposed approach, we include experiments on tabular data, images, and text; all showing improved explanations. In particular, MeLIME generated more meaningful explanations on the MNIST dataset than methods such as GuidedBackprop, SmoothGrad, and Layer-wise Relevance Propagation. MeLIME is available on https://github.com/tiagobotari/melime.

cs.LG↗

arXiv2020

Predicting and interpreting oxide glass properties by machine learning using large datasets

With the advent of powerful computer simulation techniques, it is time to move from the widely used knowledge-guided empirical methods to approaches driven by data science, mainly machine learning algorithms. We investigated the predictive performance of three machine learning algorithms for six different glass properties. For such, we used an extensive dataset of about 150,000 oxide glasses, which was segmented into smaller datasets for each property investigated. Using the decision tree induction, k-nearest neighbors, and random forest algorithms, selected from a previous study of six algorithms, we induced predictive models for glass transition temperature, liquidus temperature, elastic modulus, thermal expansion coefficient, refractive index, and Abbe number. Moreover, each model was induced with default and tuned hyperparameter values. We demonstrate that, apart from the elastic modulus (which had the smallest training dataset), the induced predictive models for the other five properties yield a comparable uncertainty to the usual data spread. However, for glasses with extremely low or high values of these properties, the prediction uncertainty is significantly higher. Finally, as expected, glasses containing chemical elements that are poorly represented in the training set yielded higher prediction errors. The method developed here calls attention to the success and possible pitfalls of machine learning algorithms. The analysis of the SHAP values indicated the key elements that increase or decrease the value of the modeled properties. It also estimated the maximum possible increase or decrease. Insights gained by this analysis can help empirical compositional tuning and computer-aided inverse design of glass formulations.

cond-mat.mtrl-sci↗

arXiv2020

Similarity of Precursors in Solid-state Synthesis as Text-Mined from Scientific Literature

Collecting and analyzing the vast amount of information available in the solid-state chemistry literature may accelerate our understanding of materials synthesis. However, one major problem is the difficulty of identifying which materials from a synthesis paragraph are precursors or are target materials. In this study, we developed a two-step Chemical Named Entity Recognition (CNER) model to identify precursors and targets, based on information from the context around material entities. Using the extracted data, we conducted a meta-analysis to study the similarities and differences between precursors in the context of solid-state synthesis. To quantify precursor similarity, we built a substitution model to calculate the viability of substituting one precursor with another while retaining the target. From a hierarchical clustering of the precursors, we demonstrate that "chemical similarity" of precursors can be extracted from text data. Quantifying the similarity of precursors helps provide a foundation for suggesting candidate reactants in a predictive synthesis model.

cond-mat.mtrl-sci↗

arXiv2019

NLS: an accurate and yet easy-to-interpret regression method

An important feature of successful supervised machine learning applications is to be able to explain the predictions given by the regression or classification model being used. However, most state-of-the-art models that have good predictive power lead to predictions that are hard to interpret. Thus, several model-agnostic interpreters have been developed recently as a way of explaining black-box classifiers. In practice, using these methods is a slow process because a novel fitting is required for each new testing instance, and several non-trivial choices must be made. We develop NLS (neural local smoother), a method that is complex enough to give good predictions, and yet gives solutions that are easy to be interpreted without the need of using a separate interpreter. The key idea is to use a neural network that imposes a local linear shape to the output layer. We show that NLS leads to predictive power that is comparable to state-of-the-art machine learning models, and yet is easier to interpret.

stat.ML↗

arXiv2019

Local Interpretation Methods to Machine Learning Using the Domain of the Feature Space

As machine learning becomes an important part of many real world applications affecting human lives, new requirements, besides high predictive accuracy, become important. One important requirement is transparency, which has been associated with model interpretability. Many machine learning algorithms induce models difficult to interpret, named black box. Moreover, people have difficulty to trust models that cannot be explained. In particular for machine learning, many groups are investigating new methods able to explain black box models. These methods usually look inside the black models to explain their inner work. By doing so, they allow the interpretation of the decision making process used by black box models. Among the recently proposed model interpretation methods, there is a group, named local estimators, which are designed to explain how the label of particular instance is predicted. For such, they induce interpretable models on the neighborhood of the instance to be explained. Local estimators have been successfully used to explain specific predictions. Although they provide some degree of model interpretability, it is still not clear what is the best way to implement and apply them. Open questions include: how to best define the neighborhood of an instance? How to control the trade-off between the accuracy of the interpretation method and its interpretability? How to make the obtained solution robust to small variations on the instance to be explained? To answer to these questions, we propose and investigate two strategies: (i) using data instance properties to provide improved explanations, and (ii) making sure that the neighborhood of an instance is properly defined by taking the geometry of the domain of the feature space into account. We evaluate these strategies in a regression task and present experimental results that show that they can improve local explanations.

cs.LG↗

arXiv2018

Mechanical Properties of Schwarzites - A Fully Atomistic Reactive Molecular Dynamics Investigation

Schwarzites are crystalline, 3D porous structures with stable negative curvature formed of sp2-hybridized carbon atoms. These structures present topologies with tunable porous size and shape and unusual mechanical properties. In this work, we have investigated the mechanical behavior under compressive strains and energy absorption of four different Schwarzites, through reactive molecular dynamics simulations, using the ReaxFF force field as available in the LAMMPS code. We considered two Schwarzites families, the so-called Gyroid and Primitive and two structures from each family. Our results also show they exhibit remarkable resilience under mechanical compression. They can be reduced to half of their original size before structural failure (fracture) occurs.

cond-mat.mes-hall↗

arXiv2017

Mechanical Properties and Fracture Patterns of Graphene (Graphitic) Nanowiggles

Graphene nanowiggles (GNW) are graphene-based nanostructures obtained by making alternated regular cuts in pristine graphene nanoribbons. GNW were recently synthesized and it was demonstrated that they exhibit tunable electronic and magnetic properties by just varying their shape. Here, we have investigated the mechanical properties and fracture patterns of a large number of GNW of different shapes and sizes using fully atomistic reactive molecular dynamics simulations. Our results show that the GNW mechanical properties are strongly dependent on its shape and size and, as a general trend narrow sheets have larger ultimate strength and Young's modulus than wide ones. The estimated Young's modulus values were found to be in a range of ~ 100-1000 GPa and the ultimate strength in a range of ~ 20-110 GPa, depending on GNW shape. Also, super-ductile behaviour under strain was observed for some structures.

cond-mat.mtrl-sci↗

arXiv2016

Towards rational design of carbon nitride photocatalysts: Identification of cyanamide "defects" as catalytically relevant sites

The heptazine-based polymer melon (also known as graphitic carbon nitride, g-C3N4), is a promising photocatalyst for hydrogen evolution. Nonetheless, attempts to improve its inherently low activity are rarely based on rational approaches due to a lack of fundamental understanding of its mechanistic operation. Here, we employ molecular heptazine-based model catalysts to identify the cyanamide moiety as a photocatalytically relevant "defect". We exploit this knowledge for the rational design of a carbon nitride polymer populated with cyanamide groups, yielding a material with 12- and 16-times the hydrogen evolution rate and apparent quantum efficiency (400 nm), respectively, compared to the benchmark melon. Computational modelling and material characterization suggest this moiety improves co-ordination (and, in turn, charge transfer kinetics) to the platinum co-catalyst and enhances the separation of the photo-generated charge carriers. The demonstrated knowledge transfer for rational catalyst design presented here provides the conceptual framework for engineering high performance heptazine-based photocatalysts.

cond-mat.mtrl-sci↗

arXiv2016

Graphene healing mechanisms: A theoretical investigation

Large holes in graphene membranes were recently shown to heal, either at room temperature during a low energy STEM experiment, or by annealing at high temperatures. However, the details of the healing mechanism remain unclear. We carried out fully atomistic reactive molecular dynamics simulations in order to address these mechanisms under different experimental conditions. Our results show that, if a carbon atom source is present, high temperatures can provide enough energy for the carbon atoms to overcome the potential energy barrier and to produce perfect reconstruction of the graphene hexagonal structure. At room temperature, this perfect healing is only possible if the heat effects of the electron beam from STEM experiment are explicitly taken into account. The reconstruction process of a perfect or near perfect graphene structure involves the formation of linear carbon chains, as well as rings containing 5, 6, 7 and 8 atoms with planar (Stone-Wales) and non-planar (lump like) structures. These results shed light on the healing mechanism of graphene when subjected to different experimental conditions. Additionally, the methodology presented here can be useful for investigating the tailoring and manipulations of other nano-structures.

cond-mat.mes-hall↗

arXiv2014

One-dimensional Silicon and Germanium Nanostructures With No Carbon Analogues

In this work we report new silicon and germanium tubular nanostructures with no corresponding stable carbon analogues. The electronic and mechanical properties of these new tubes were investigated through ab initio methods. Our results show that the structures have lower energy than their corresponding nanoribbon structures and are stable up to high temperatures (500 and 1000 K, for silicon and germanium tubes, respectively). Both tubes are semiconducting with small indirect band gaps, which can be significantly altered by both compressive and tensile strains. Large bandgap variations of almost 50% were observed for strain rates as small as 3%, suggesting possible applications in sensor devices. They also present high Young's modulus values (0.25 and 0.15 TPa, respectively). TEM images were simulated to help the identification of these new structures.

cond-mat.mes-hall↗

arXiv2012

A one-dimensional Fermi accelerator model with moving wall described by a nonlinear van der Pol oscillator

A modification of the one-dimensional Fermi accelerator model is considered in this work. The dynamics of a classical particle of mass $m$, confined to bounce elastically between two rigid walls where one is described by a non-linear van der Pol type oscillator while the other one is fixed, working as a re-injection mechanism of the particle for a next collision, is carefully made by the use of a two-dimensional non-linear mapping. Two cases are considered: (i) the situation where the particle has mass negligible as compared to the mass of the moving wall and does not affect the motion of it; (ii) the case where collisions of the particle does affect the movement of the moving wall. For case (i) the phase space is of mixed type leading us to observe a scaling of the average velocity as a function of the parameter ($\c{hi}$) controlling the non-linearity of the moving wall. For large $\c{hi}$, a diffusion on the velocity is observed leading us to conclude that Fermi acceleration is taking place. On the other hand for case (ii), the motion of the moving wall is affected by collisions with the particle. However due to the properties of the van der Pol oscillation, the moving wall relaxes again to a limit cycle. Such kind of motion absorbs part of the energy of the particle leading to a suppression of the unlimited energy gain as observed in case (i). The phase space shows a set of attractors of different periods whose basin of attraction has a complicate organization.

nlin.CD↗