Search

Africa faces a huge shortage of dermatologists, with less than one per million people. This is in stark contrast to the high demand for dermatologic care, with 80% of the paediatric population suffering from largely untreated skin conditions. The integration of AI into healthcare sparks significant hope for treatment accessibility, especially through the development of AI-supported teledermatology. Current AI models are predominantly trained on white-skinned patients and do not generalize well enough to pigmented patients. The PASSION project aims to address this issue by collecting images of skin diseases in Sub-Saharan countries with the aim of open-sourcing this data. This dataset is the first of its kind, consisting of 1,653 patients for a total of 4,901 images. The images are representative of telemedicine settings and encompass the most common paediatric conditions: eczema, fungals, scabies, and impetigo. We also provide a baseline machine learning model trained on the dataset and a detailed performance analysis for the subpopulations represented in the dataset. The project website can be found at https://passionderm.github.io/., Comment: MICCAI 2024

The Time-Domain And MultiMessenger (TDAMM) Communications Science Analysis Group (TDAMMCommSAG) was formulated to describe the unique technical challenges of communicating rapidly to and from NASA astrophysics missions studying the most variable, transient, and extreme objects in the Universe. This report describes the study of if and how the transition from current NASA-operated space and ground relays to commercial services will adequately serve these missions. Depending on the individual mission requirements and Concept of Operations (ConOps), TDAMM missions may utilize a rapid low-rate demand access service, a low-rate continuous contact service, low-latency downlink upon demand, or a higher-latency but regular relay service. The specific implementations can vary via space relay or direct to Earth, but requires flexibility and adaptability using modern software infrastructure. The study team reviewed the current state of NASA communications services and future commercial and NASA communications services under study and in development. We explored the communications capabilities driving from the behavior of the astrophysical objects themselves., Comment: 20 page, 1 figure

Violence and abnormal behavior detection research have known an increase of interest in recent years, due mainly to a rise in crimes in large cities worldwide. In this work, we propose a deep learning architecture for violence detection which combines both recurrent neural networks (RNNs) and 2-dimensional convolutional neural networks (2D CNN). In addition to video frames, we use optical flow computed using the captured sequences. CNN extracts spatial characteristics in each frame, while RNN extracts temporal characteristics. The use of optical flow allows to encode the movements in the scenes. The proposed approaches reach the same level as the state-of-the-art techniques and sometime surpass them. It was validated on 3 databases achieving good results., Comment: 11 pages, 7 figures, 2020 IEEE International Conference on Systems, Man, and Cybernetics (SMC)

Abnormal behavior detection, action recognition, fight and violence detection in videos is an area that has attracted a lot of interest in recent years. In this work, we propose an architecture that combines a Bidirectional Gated Recurrent Unit (BiGRU) and a 2D Convolutional Neural Network (CNN) to detect violence in video sequences. A CNN is used to extract spatial characteristics from each frame, while the BiGRU extracts temporal and local motion characteristics using CNN extracted features from multiple frames. The proposed end-to-end deep learning network is tested in three public datasets with varying scene complexities. The proposed network achieves accuracies up to 98%. The obtained results are promising and show the performance of the proposed end-to-end approach., Comment: 8 pages, 6 figures, 2020 International Conference on Image Analysis and Recognition (ICIAR)

This work explores the antimicrobial efficacy of ZnO nanoparticles (NPs) films with varying microstructure, crystallographic orientation and surface roughness. Thin films with different microstructures were obtained by varying precursor concentrations (0.1 M, 0.25 M, and 0.4 M) and solvent types (Ethanol, Propanol, and Butanol). The characterization of the films by X-ray diffraction (XRD) revealed films with varying degrees of orientation along the c-axis, which increased with the concentration of the precursor and with the decrease in the carbon chain length of the alcohol. Raman spectroscopy analysis confirmed the formation of a hexagonal wurtzite ZnO. Scanning Electron Microscopy (SEM) analysis revealed the formation of dense and homogeneous films with a thickness of 93 nm. Atomic force microscopy (AFM) showed changes in surface irregularity and an increase in surface roughness with the carbon length of solvent and a maximum roughness at a precursor concentration of 0.25 M in propanol. The antimicrobial activity was assessed following ISO 22196:2011 standards with few modifications. Results indicated that ZnO NPs films with rougher surfaces reduced coli (gram(-)), Pseudomonas aeruginosa (gram(-)) and Staphylococcus aureus (gram(+) ) bacteria populations due to an increase of contact surface between NPs with bacteria. Their effectiveness against Candida albicans was moderate, with a maximum reduction of 0.72 Log10 CFU/cm$^2$ with rougher film.

Using GPU-accelerated state-vector emulation, we propose to embed a quantum computing ansatz into density-functional theory via density-based basis-set corrections (DBBSC) to obtain quantitative quantum-chemistry results on molecules that would otherwise require brute-force quantum calculations using hundreds of logical qubits. Indeed, accessing a quantitative description of chemical systems while minimizing quantum resources is an essential challenge given the limited qubit capabilities of current quantum processors. We provide a shortcut towards chemically accurate quantum computations by approaching the complete-basis-set limit through coupling the DBBSC approach, applied to any given variational ansatz, to an on-the-fly crafting of basis sets specifically adapted to a given system and user-defined qubit budget. The resulting approach self-consistently accelerates the basis-set convergence, improving electronic densities, ground-state energies, and first-order properties (e.g. dipole moments), but can also serve as a classical, a posteriori, energy correction to quantum hardware calculations with expected applications in drug design and materials science.

Effective communication, specifically through documentation, is the beating heart of collaboration among contributors in software development. Recent advancements in language models (LMs) have enabled the introduction of a new type of actor in that ecosystem: LM-powered assistants capable of code generation, optimization, and maintenance. Our study investigates the efficacy of small language models (SLMs) for generating high-quality docstrings by assessing accuracy, conciseness, and clarity, benchmarking performance quantitatively through mathematical formulas and qualitatively through human evaluation using Likert scale. Further, we introduce DocuMint, as a large-scale supervised fine-tuning dataset with 100,000 samples. In quantitative experiments, Llama 3 8B achieved the best performance across all metrics, with conciseness and clarity scores of 0.605 and 64.88, respectively. However, under human evaluation, CodeGemma 7B achieved the highest overall score with an average of 8.3 out of 10 across all metrics. Fine-tuning the CodeGemma 2B model using the DocuMint dataset led to significant improvements in performance across all metrics, with gains of up to 22.5% in conciseness. The fine-tuned model and the dataset can be found in HuggingFace and the code can be found in the repository., Comment: 12 pages, 4 figures

Nosocomial infections (NIs) are prevalent in intensive care units due to antibiotic overuse. Metal oxide nanoparticles (NPs), like ZnO, offer potential solutions, yet understanding how NPs size impacts their antibacterial efficacy are lacking. This study focuses on the effect of nanoparticle size on kinetics of bacterial strains growth. NPs were synthesized using a sol-gel process with monoethanolamine (MEA) and water, characterized using X-ray diffraction (XRD), transmission electron microscopy (TEM), and Raman spectroscopy, confirming crystallization and size variations. ZnO NPs with mean size of 22, 35 and 66 nm were used against the most common nosocomial bacteria strains Escherichia coli (Gram-negative), Pseudomonas aeruginosa (Gram-negative), and Staphylococcus aureus (Gram-positive). The evaluation of NPs minimal inhibitory concentration (MIC) and bactericidal concentration (MBC) revealed superior antibacterial activity in smaller NPs. The bacterial population was monitored via optical absorbance, showing reduced specific growth rate, prolonged latency period, and increased inhibition percentage with smaller NPs, indicating a substantial deceleration in the growth of microorganisms. Pseudomonas aeruginosa exhibited the smallest sensitivity to ZnO NPs, attributed to its environmental stress resistance. Furthermore, the antibacterial efficacy of paint containing 1 wt% of 22 nm ZnO NPs was assessed and displayed activity against E. coli and S. aureus.

We present the first application to real molecular systems of the recently proposed linear-response theory for the density-based basis-set correction method [J. Chem. Phys. 158, 234107 (2023)]. We apply this approach to accelerate the basis-set convergence of excitation energies in the equation-of-motion coupled-cluster singles doubles (EOM-CCSD) method. We use an approximate linear-response framework which neglects the second-order derivative of the basis-set correction density functional and consists in simply adding to the usual Hamiltonian the one-electron potential generated by the first-order derivative of the functional. This additional basis-set correction potential is evaluated at the Hartree-Fock density, leading to a very computationally cheap basis-set correction. We tested this approach over a set of about 30 excitation energies computed for five small molecular systems and found that the excitation energies from the ground state to Rydberg states are the main source of basis-set error. These excitation energies systematically increase when the size of the basis set is increased, suggesting a biased description in favour of the excited state. Despite the simplicity of the present approach, the results obtained with the basis-set corrected EOM-CCSD method are encouraging as they yield to a mean absolute deviation of 0.02 eV for the aug-cc-pVTZ basis set, while it is of 0.04 eV using the standard EOM-CCSD method. This might open the path to an alternative to explicitly correlated approaches to accelerate the basis-set convergence of excitation energies.

In the last 5 years, the availability of large audio datasets in African countries has opened unlimited opportunities to build machine intelligence (MI) technologies that are closer to the people and speak, learn, understand, and do businesses in local languages, including for those who cannot read and write. Unfortunately, these audio datasets are not fully exploited by current MI tools, leaving several Africans out of MI business opportunities. Additionally, many state-of-the-art MI models are not culture-aware, and the ethics of their adoption indexes are questionable. The lack thereof is a major drawback in many applications in Africa. This paper summarizes recent developments in machine intelligence in Africa from a multi-layer multiscale and culture-aware ethics perspective, showcasing MI use cases in 54 African countries through 400 articles on MI research, industry, government actions, as well as uses in art, music, the informal economy, and small businesses in Africa. The survey also opens discussions on the reliability of MI rankings and indexes in the African continent as well as algorithmic definitions of unclear terms used in MI., Comment: Accepted and to be presented at DSAI 2024

We present an enhanced algorithm for exploring mirror symmetry for elliptic curves through the correspondence of algebraic and tropical geometry, focusing on Gromov-Witten invariants of elliptic curves and, in particular, Hurwitz numbers. We present a new highly efficient algorithm for computing generating series for these numbers. We have implemented the algorithm both using Singular and OSCAR. The implementations outperform by far the current method provided in Singular. The OSCAR implementation, benefiting in particular from just-in-time compilation, again by far outperforms the implementation of the new algorithm in Singular. This advancement in computing the Gromov-Witten invariants facilitates a study of number theoretic and geometric properties of the generating series, including quasi-modularity and homogeneity., Comment: 20 pages, 5 figures, 2 tables. arXiv admin note: text overlap with arXiv:1309.5893

Recently, machine learning of the branch and bound algorithm has shown promise in approximating competent solutions to NP-hard problems. In this paper, we utilize and comprehensively compare the outcomes of three neural networks--graph convolutional neural network (GCNN), GraphSAGE, and graph attention network (GAT)--to solve the capacitated vehicle routing problem. We train these neural networks to emulate the decision-making process of the computationally expensive Strong Branching strategy. The neural networks are trained on six instances with distinct topologies from the CVRPLIB and evaluated on eight additional instances. Moreover, we reduced the minimum number of vehicles required to solve a CVRP instance to a bin-packing problem, which was addressed in a similar manner. Through rigorous experimentation, we found that this approach can match or improve upon the performance of the branch and bound algorithm with the Strong Branching strategy while requiring significantly less computational time. The source code that corresponds to our research findings and methodology is readily accessible and available for reference at the following web address: https://isotlaboratory.github.io/ml4vrp

In the context of finite sums minimization, variance reduction techniques are widely used to improve the performance of state-of-the-art stochastic gradient methods. Their practical impact is clear, as well as their theoretical properties. Stochastic proximal point algorithms have been studied as an alternative to stochastic gradient algorithms since they are more stable with respect to the choice of the step size. However, their variance-reduced versions are not as well studied as the gradient ones. In this work, we propose the first unified study of variance reduction techniques for stochastic proximal point algorithms. We introduce a generic stochastic proximal-based algorithm that can be specified to give the proximal version of SVRG, SAGA, and some of their variants. For this algorithm, in the smooth setting, we provide several convergence rates for the iterates and the objective function values, which are faster than those of the vanilla stochastic proximal point algorithm. More specifically, for convex functions, we prove a sublinear convergence rate of $O(1/k)$. In addition, under the Polyak-{\L}ojasiewicz (PL) condition, we obtain linear convergence rates. Finally, our numerical experiments demonstrate the advantages of the proximal variance reduction methods over their gradient counterparts in terms of the stability with respect to the choice of the step size in most cases, especially for difficult problems.

The basis-set correction method based on density-functional theory consists in correcting the energy calculated by a wave-function method with a given basis set by a density functional. This basis-set correction density functional incorporates the short-range electron correlation effects missing in the basis set. This results in accelerated basis convergences of ground-state energies to the complete-basis-set limit. In this work, we extend the basis-set correction method to a linear-response formalism for calculating excited-state energies. We give the general linear-response equations, as well as the more specific equations for configuration-interaction wave functions. As a proof of concept, we apply this approach to the calculations of excited-state energies in a one-dimensional two-electron model system with harmonic potential and a Dirac-delta electron-electron interaction. The results obtained with full-configuration-interaction wave functions expanded in a basis of Hermite functions and a local-density-approximation basis-set correction functional show that the present approach does not help in accelerating the basis convergence of excitation energies. However, we show that it significantly accelerates basis convergences of excited-state total energies.

In this paper, we present MasakhaPOS, the largest part-of-speech (POS) dataset for 20 typologically diverse African languages. We discuss the challenges in annotating POS for these languages using the UD (universal dependencies) guidelines. We conducted extensive POS baseline experiments using conditional random field and several multilingual pre-trained language models. We applied various cross-lingual transfer models trained with data available in UD. Evaluating on the MasakhaPOS dataset, we show that choosing the best transfer language(s) in both single-source and multi-source setups greatly improves the POS tagging performance of the target languages, in particular when combined with cross-lingual parameter-efficient fine-tuning methods. Crucially, transferring knowledge from a language that matches the language family and morphosyntactic properties seems more effective for POS tagging in unseen languages., Comment: Accepted to ACL 2023 (Main conference)