Search

This research introduces a Machine Learning-centric approach to replicate olfactory experiences, validated through experimental quantification of perfume perception. Key contributions encompass a hybrid model connecting perfume molecular structure to human olfactory perception. This model includes an AI-driven molecule generator (utilizing Graph and Generative Neural Networks), quantification and prediction of odor intensity, and refinery of optimal solvent and molecule combinations for desired fragrances. Additionally, a thermodynamic-based model establishes a link between olfactory perception and liquid-phase concentrations. The methodology employs Transfer Learning and selects the most suitable molecules based on vapor pressure and fragrance notes. Ultimately, a mathematical optimization problem is formulated to minimize discrepancies between new and target olfactory experiences. The methodology is validated by reproducing two distinct olfactory experiences using available experimental data.

Dynamical masses of young planets aged between 10 and 200 Myr detected in imaging play a crucial role in shaping models of giant planet formation. Regrettably, only a few such objects possess these characteristics. Furthermore, the evolutionary pattern of young sub-stellar companions in near-infrared colour-magnitude diagrams might diverge from free-floating objects, possibly due to differing formation processes. The recent identification of a giant planet around AF Lep, part of the beta Pic moving group (BPMG), encouraged us to re-examine these points. We considered updated dynamical masses and luminosities for the sub-stellar objects in the BPMG. In addition, we compared the properties of sub-stellar companions and free-floating objects in the BPMG and other young associations remapping the positions of the objects in the colour-magnitude diagram into a dustiness-temperature plane. We found that cold-start evolutionary models do not reproduce the mass-luminosity relation for sub-stellar companions in the BPMG. This aligns rather closely with predictions from 'hot start' scenarios and is consistent with recent planet formation models. We obtain rather good agreement with masses from photometry and the remapping approach compared to actual dynamical masses. We also found a strong suggestion that the near-infrared colour-magnitude diagram for young companions is different from that of free-floating objects belonging to the same young associations. If confirmed by further data, this last result would imply that cloud settling - which likely causes the transition between L and T spectral type - occurs at a lower effective temperature in young companions than in free-floating objects. This might tentatively be explained with a different chemical composition., Comment: Accepted on Astronomy and Astrophysics. 19 pages, 12 figures

Objective: In cochlear implant users with residual acoustic hearing, compound action potentials (CAPs) can be evoked by acoustic (aCAP) or electric (eCAP) stimulation and recorded through the electrodes of the implant. We propose a novel computational model to simulate aCAPs and eCAPs in humans, considering the interaction between combined electric-acoustic stimulation that occurs in the auditory nerve. Methods: The model consists of three components: a 3D finite element method model of an implanted cochlea, a phenomenological single-neuron spiking model for electric-acoustic stimulation, and a physiological multi-compartment neuron model to simulate the individual nerve fiber contributions to the CAP. Results: The CAP morphologies closely resembled those known from humans. The spread of excitation derived from eCAPs by varying the recording electrode along the cochlear implant electrode array was consistent with published human data. The predicted CAP amplitude growth functions largely resembled human data, with deviations in absolute CAP amplitudes for acoustic stimulation. The model reproduced the suppression of eCAPs by simultaneously presented acoustic tone bursts for different masker frequencies and probe stimulation electrodes. Conclusion: The proposed model can simulate CAP responses to electric, acoustic, or combined electric-acoustic stimulation. It considers the dependence on stimulation and recording sites in the cochlea, as well as the interaction between electric and acoustic stimulation in the auditory nerve. Significance: The model enhances comprehension of CAPs and peripheral electric-acoustic interaction. It can be used in the future to investigate objective methods, such as hearing threshold assessment or estimation of neural health through aCAPs or eCAPs., Comment: This work has been submitted to the IEEE for possible publication

Language models are now capable of solving tasks that require dealing with long sequences consisting of hundreds of thousands of tokens. However, they often fail on tasks that require repetitive use of simple rules, even on sequences that are much shorter than those seen during training. For example, state-of-the-art LLMs can find common items in two lists with up to 20 items but fail when lists have 80 items. In this paper, we introduce Lissard, a benchmark comprising seven tasks whose goal is to assess the ability of models to process and generate wide-range sequence lengths, requiring repetitive procedural execution. Our evaluation of open-source (Mistral-7B and Mixtral-8x7B) and proprietary models (GPT-3.5 and GPT-4) show a consistent decline in performance across all models as the complexity of the sequence increases. The datasets and code are available at https://github.com/unicamp-dl/Lissard

ExaRanker recently introduced an approach to training information retrieval (IR) models, incorporating natural language explanations as additional labels. The method addresses the challenge of limited labeled examples, leading to improvements in the effectiveness of IR models. However, the initial results were based on proprietary language models such as GPT-3.5, which posed constraints on dataset size due to its cost and data privacy. In this paper, we introduce ExaRanker-Open, where we adapt and explore the use of open-source language models to generate explanations. The method has been tested using different LLMs and datasets sizes to better comprehend the effective contribution of data augmentation. Our findings reveal that incorporating explanations consistently enhances neural rankers, with benefits escalating as the LLM size increases. Notably, the data augmentation method proves advantageous even with large datasets, as evidenced by ExaRanker surpassing the target baseline by 0.6 nDCG@10 points in our study. To encourage further advancements by the research community, we have open-sourced both the code and datasets at https://github.com/unicamp-dl/ExaRanker.

This report introduces LLaMandement, a state-of-the-art Large Language Model, fine-tuned by the French government and designed to enhance the efficiency and efficacy of processing parliamentary sessions (including the production of bench memoranda and documents required for interministerial meetings) by generating neutral summaries of legislative proposals. Addressing the administrative challenges of manually processing a growing volume of legislative amendments, LLaMandement stands as a significant legal technological milestone, providing a solution that exceeds the scalability of traditional human efforts while matching the robustness of a specialized legal drafter. We release all our fine-tuned models and training data to the community., Comment: 21 pages, 9 figures

Predictive modeling of CO2 storage sites requires a detailed understanding of physico-chemical processes and potential challenges for scale-up. Dramatic injectivity decline may occur due to salt precipitation pore clogging in high-salinity reservoirs, even over a short time frame. This study aims to elucidate the adverse impact of CO2-induced salt crystallization in porous media on the geomechanical properties of near-wellbore reservoir sandstones. As the impact of salt precipitation cannot be isolated from the precursor effects of interaction with CO2 and carbonic acid, we initiated our study by a comprehensive review of CO2 chemo-mechanical interactions with sandstones. We conducted laboratory geochemical CO2-brine-rock interactions at elevated pressures and temperatures on two sets of porous sandstone with contrasting petrophysical qualities. Two paths were followed: treatment with (a) CO2-acidified brine at 10 MPa fluid pressure and 60C for 7 days, and a second subset continuation with (b) supercritical injection until complete dry-out and salt precipitation. Afterward, the core samples were tested in a triaxial apparatus at varying stresses and temperatures. The elastic moduli of intact, CO2-reacted, and salt-damaged sandstones were juxtaposed to elucidate the extent of crystallization damages. The salt-affected specimens showed a maximum of 50 percent reduction in Young's and shear moduli and twice an increase in Poisson's ratio compared to intact condition. The deterioration was notably higher for the tighter rocks with higher initial stiffness.

This paper introduces INACIA (Instru\c{c}\~ao Assistida com Intelig\^encia Artificial), a groundbreaking system designed to integrate Large Language Models (LLMs) into the operational framework of Brazilian Federal Court of Accounts (TCU). The system automates various stages of case analysis, including basic information extraction, admissibility examination, Periculum in mora and Fumus boni iuris analyses, and recommendations generation. Through a series of experiments, we demonstrate INACIA's potential in extracting relevant information from case documents, evaluating its legal plausibility, and formulating propositions for judicial decision-making. Utilizing a validation dataset alongside LLMs, our evaluation methodology presents a novel approach to assessing system performance, correlating highly with human judgment. These results underscore INACIA's potential in complex legal task handling while also acknowledging the current limitations. This study discusses possible improvements and the broader implications of applying AI in legal contexts, suggesting that INACIA represents a significant step towards integrating AI in legal systems globally, albeit with cautious optimism grounded in the empirical findings.

Carbon dioxide (CO2) trapping in capillary networks of reservoir rocks is a pathway to long-term geological storage. At pore scale, the CO2 trapping potential depends on injection pressure, temperature, and the rock's interaction with the surrounding fluids. Modeling this interaction requires adequate representations of both capillary volume and surface. For the lack of scalable representations, however, the prediction of a rock's CO2 storage potential has been challenging. Here, we report how to represent a rock's pore space by statistically sampled capillary networks (ssCN) that preserve morphological rock characteristics. We have used the ssCN method to simulate CO2 drainage within a representative sandstone sample at reservoir pressures and temperatures, exploring intermediate- and CO2-wet conditions. This wetting regime is often neglected, despite evidence of plausibility. By raising pressure and temperature we observe increasing CO2 penetration within the capillary network. For contact angles approaching 90 degrees, the CO2 saturation exhibits a pronounced maximum reaching 80 percent of the accessible pore volume. This is about twice as high as the saturation values reported previously. For enabling validation of our results and a broader application of our methodology, we have made available the rock tomography data, the digital rock computational workflows, and the ssCN models used in this study., Comment: Main article: 16 pages and 3 figures. Supplementary Information: 17 pages, 3 equations, 13 figures and 3 tables. Includes DOI for accompanying digital rock data, URL of github repository for flow simulator code, and of accompanying digital rock processing python code and automation of the scientific workflows

This paper presents a detailed case study examining the application of Large Language Models (LLMs) in the construction of test cases within the context of software engineering. LLMs, characterized by their advanced natural language processing capabilities, are increasingly garnering attention as tools to automate and enhance various aspects of the software development life cycle. Leveraging a case study methodology, we systematically explore the integration of LLMs in the test case construction process, aiming to shed light on their practical efficacy, challenges encountered, and implications for software quality assurance. The study encompasses the selection of a representative software application, the formulation of test case construction methodologies employing LLMs, and the subsequent evaluation of outcomes. Through a blend of qualitative and quantitative analyses, this study assesses the impact of LLMs on test case comprehensiveness, accuracy, and efficiency. Additionally, delves into challenges such as model interpretability and adaptation to diverse software contexts. The findings from this case study contributes with nuanced insights into the practical utility of LLMs in the domain of test case construction, elucidating their potential benefits and limitations. By addressing real-world scenarios and complexities, this research aims to inform software practitioners and researchers alike about the tangible implications of incorporating LLMs into the software testing landscape, fostering a more comprehensive understanding of their role in optimizing the software development process.

Transfer learning allows for resource-efficient geographic transfer of pre-trained field delineation models. However, the scarcity of labeled data for complex and dynamic smallholder landscapes, particularly in Sub-Saharan Africa, remains a major bottleneck for large-area field delineation. This study explores opportunities of using sparse field delineation pseudo labels for fine-tuning models across geographies and sensor characteristics. We build on a FracTAL ResUNet trained for crop field delineation in India (median field size of 0.24 ha) and use this pre-trained model to generate pseudo labels in Mozambique (median field size of 0.06 ha). We designed multiple pseudo label selection strategies and compared the quantities, area properties, seasonal distribution, and spatial agreement of the pseudo labels against human-annotated training labels (n = 1,512). We then used the human-annotated labels and the pseudo labels for model fine-tuning and compared predictions against human field annotations (n = 2,199). Our results indicate i) a good baseline performance of the pre-trained model in both field delineation and field size estimation, and ii) the added value of regional fine-tuning with performance improvements in nearly all experiments. Moreover, we found iii) substantial performance increases when using only pseudo labels (up to 77% of the IoU increases and 68% of the RMSE decreases obtained by human labels), and iv) additional performance increases when complementing human annotations with pseudo labels. Pseudo labels can be efficiently generated at scale and thus facilitate domain adaptation in label-scarce settings. The workflow presented here is a stepping stone for overcoming the persisting data gaps in heterogeneous smallholder agriculture of Sub-Saharan Africa, where labels are commonly scarce.

Recent developments and research in modern machine learning have led to substantial improvements in the geospatial field. Although numerous deep learning architectures and models have been proposed, the majority of them have been solely developed on benchmark datasets that lack strong real-world relevance. Furthermore, the performance of many methods has already saturated on these datasets. We argue that a shift from a model-centric view to a complementary data-centric perspective is necessary for further improvements in accuracy, generalization ability, and real impact on end-user applications. Furthermore, considering the entire machine learning cycle-from problem definition to model deployment with feedback-is crucial for enhancing machine learning models that can be reliable in unforeseen situations. This work presents a definition as well as a precise categorization and overview of automated data-centric learning approaches for geospatial data. It highlights the complementary role of data-centric learning with respect to model-centric in the larger machine learning deployment cycle. We review papers across the entire geospatial field and categorize them into different groups. A set of representative experiments shows concrete implementation examples. These examples provide concrete steps to act on geospatial data with data-centric machine learning approaches., Comment: Accepted to Geoscience and Remote Sensing Magazine

We show that the complexity of the billiard in a typical polygon grows cubically and the number of saddle connections grows quadratically along certain subsequences. It is known that the set of points whose first n-bounces hits the same sequence of sides as the orbit of an aperiodic phase point z converges to z. We establishe a polynomial lower bound estimate on this convergence rate for almost every z. This yields an upper bound on the upper metric complexity and upper slow entropy of polygonal billiards. We also prove significant deviations from the expected convergence behavior. Finally we extend these results to higher dimensions as well as to arbitrary invariant measures.

In the realm of statistical physics, the number of states in which a system can be realized with a given energy is a key concept that bridges the microscopic and macroscopic descriptions of physical systems. For quantum systems, many approaches rely on the solution of the Schr\"odinger equation. In this work, we demonstrate how the recently developed rodeo algorithm can be utilized to determine the number of states associated with all energy levels without any prior knowledge of the eigenstates. Quantum computers, with their innate ability to address the intricacies of quantum systems, make this approach particularly promising for the study of the thermodynamics of those systems. To illustrate the procedure's effectiveness, we apply it to compute the number of states of the 1D transverse-field Ising model and, consequently, its specific heat, proving the reliability of the method presented here., Comment: 6 pages, 4 figures

This study presents a methodology for surrogate optimization of cyclic adsorption processes, focusing on enhancing Pressure Swing Adsorption units for carbon dioxide ($CO_{2}$) capture. We developed and implemented a multiple-input, single-output (MISO) framework comprising two deep neural network (DNN) models, predicting key process performance indicators. These models were then integrated into an optimization framework, leveraging particle swarm optimization (PSO) and statistical analysis to generate a comprehensive Pareto front representation. This approach delineated feasible operational regions (FORs) and highlighted the spectrum of optimal decision-making scenarios. A key aspect of our methodology was the evaluation of optimization effectiveness. This was accomplished by testing decision variables derived from the Pareto front against a phenomenological model, affirming the surrogate models reliability. Subsequently, the study delved into analyzing the feasible operational domains of these decision variables. A detailed correlation map was constructed to elucidate the interplay between these variables, thereby uncovering the most impactful factors influencing process behavior. The study offers a practical, insightful operational map that aids operators in pinpointing the optimal process location and prioritizing specific operational goals.

We build a setup for path integral quantization through the Faddeev-Jackiw approach, extending it to include Grassmannian degrees of freedom, to be later implemented in a model of generalized electrodynamics that involves fourth-order derivatives in the components of a massive vector field being endowed with gauge freedom, due to an additional scalar field. Namely, the generalized Stueckelberg electrodynamics. In the first instance, we work on the free case to gain some familiarity with the program and, subsequently, we add the interaction with fermionic matter fields to complete our aim. In addition to deriving the correct classical brackets for such a model, we get the full expression for the associated generating functional and its associated integration measure.