Google Colab Research Articles

ElasticNotebook: Enabling Live Migration for Computational Notebooks

Computational notebooks (e.g., Jupyter, Google Colab) are widely used for interactive data science and machine learning. In those frameworks, users can start a session , then execute cells (i.e., a set of statements) to create variables, train models, visualize results, etc. Unfortunately, existing notebook systems do not offer live migration: when a notebook launches on a new machine, it loses its state , preventing users from continuing their tasks from where they had left off. This is because, unlike DBMS, the sessions directly rely on underlying kernels (e.g., Python/R interpreters) without an additional data management layer. Existing techniques for preserving states, such as copying all variables or OS-level checkpointing, are unreliable (often fail), inefficient, and platform-dependent. Also, re-running code from scratch can be highly time-consuming. In this paper, we introduce a new notebook system, Elastic-Notebook, that offers live migration via checkpointing/restoration using a novel mechanism that is reliable, efficient, and platform-independent. Specifically, by observing all cell executions via transparent, lightweight monitoring, ElasticNotebook can find a reliable and efficient way (i.e., replication plan ) for reconstructing the original session state, considering variable-cell dependencies, observed runtime, variable sizes, etc. To this end, our new graph-based optimization problem finds how to reconstruct all variables (efficiently) from a subset of variables that can be transferred across machines. We show that ElasticNotebook reduces end-to-end migration and restoration times by 85%-98% and 94%-99%, respectively, on a variety (i.e., Kaggle, JWST, and Tutorial) of notebooks with negligible runtime and memory overheads of <2.5% and <10%.

MONITORING AND ANALYZING DIABETIC FOOT ULCER USING RESNET50 IN DEEP LEARNING

This paper introduces a deep learning approach using the ResNet50 algorithm within Google Colab to detect diabetic foot ulcers. Early identification of these ulcers is crucial to prevent leg amputations. Leveraging a diverse dataset, the model exhibits high accuracy and sensitivity, offering a cost-effective solution for timely intervention in diabetic foot ulcers, there by improving patient outcomes.

Analisa Algoritma Naïve Bayes Classifier (NBC) Untuk Prediksi Penjualan Alat Kesehatan

The application of Data Mining in the business scope can be found in the use of Customer Relationship Management (CRM). CRM is a company's effort to manage its sales and customers more optimally. Company Sales Data can be processed into knowledge that can be used to optimize marketing strategies. Purna Karya Scientific is a company engaged in the field of medical/medical devices, laboratory equipment, chemical and dental materials as well as educational aids. In this study has used sales data at PT. After Scientific Work with attributes item code, relation, number of items, and label as class. Then classify medical device sales data by implementing the Naïve Bayes Classifier (NBC) algorithm which can predict sales results by displaying an accuracy value. Implementation was carried out using Google Colab to obtain an accuracy value of 95%, a recall value of 95%, and a precision value of 81%. The results of data on sales of medical devices with 2 classes namely "Selling" and "Not Selling". The resulting value is very good and can be used as a basis for classifying sales of medical devices by analyzing the stock of goods at PT. Full Scientific Work.

MEGADOCK-on-Colab: an easy-to-use protein–protein docking tool on Google Colaboratory

MotivationSince the advent of ColabFold, numerous software packages have been provided with Google Colaboratory-compatible ipynb files, allowing users to effortlessly test and reproduce results without the need for local installation or configuration. MEGADOCK, a protein–protein docking tool, is particularly well-suited for Google Colaboratory due to its lightweight computations and GPU acceleration capabilities. To increase accessibility and promote widespread use, it is crucial to provide a computing environment compatible with Google Colaboratory.ResultsIn this study, we report the development of a Google Colaboratory environment for running our protein–protein docking software, MEGADOCK. We provide a comprehensive ipynb file, including the compilation of MEGADOCK with the FFTW library installation on Colaboratory, the introduction of related tools using PyPI/apt, and the execution and visualization of docking structures. This streamlined environment enables users to visualize docking structures with just one click. The code is available under a CC-BY NC 4.0 license from https://github.com/ohuelab/MEGADOCK-on-Colab.

Comparison of C4.5 and Naive Bayes for Predicting Student Graduation Using Machine Learning Algorithms

Student graduation is a very important element for universities because it relates to college accreditation assessment. One of them is at the Faculty of Engineering Nurul Jadid University, which has problems completing the study period within a predetermined time. So that it can be detrimental because accreditation is less than optimal, and the number of active students makes it less ideal in teaching and learning activities. This study aimed to compare the level of accuracy using the C4.5 algorithm and Naïve Bayes method in predicting graduation on time. The C4.5 and Naïve Bayes algorithms are one of the methods in the algorithm for classifying. Tests were carried out using the C4.5 and Naïve Bayes algorithms using Google Colab with Python programming language, then validated using 10-fold cross-validation. The results of this study indicate that the Naïve Bayes method has a higher accuracy value with an accuracy rate of 96.12%, while the C4.5 algorithm method is 93.82%.

대학생의 다전공 인식과 개선 및 요구사항 분석을 통한 향후 운영 방안 연구

Objectives The purpose of this study is to present future strategies for the stable operation of a multi-major curriculum
 through the analysis of perceptions, improvements, and requirements related to multi-major.
 Methods The subjects of the study were 1,218 students at A University, which has a multi-major system in Seoul,
 and the difference was verified with the SPSS 23.0 program. For 232 people who responded to open questions,
 morphology analysis was conducted using KoNLPy based on Google Colab and then frequency analysis and
 Wordcloud were conducted.
 Results The main results are as follows. First, as a result of verifying the difference between multi-major perception
 and multi-major type classification, women were higher than men, and second, third and fourth graders were
 higher than first graders, and third and fourth graders were higher then second graders. In the case of the college,
 there was a statistically significant difference, but there was no statistically significant difference as a result of
 Scheffé's post-test. Second, as a result of verifying the difference between the preferred multi-majors, female
 students prefer double majors and male students prefer minor and convergence majors. There was no statistically
 significant difference in grades. College of humanities preferred double majors, engineering and IT colleges preferred
 minor majors, and IT preferred convergence majors. Third, as a result of verifying the difference in multi-
 major improvements, female students had high opinions on adjustment of the number of selected people by
 department,expansion of division by subject, and opportunity to participate in special activities by department.,
 and male students had high opinions on ‘strengthening academic support(professors, mentoring, teaching assistants).
 Compared to second, third, and fourth graders, first graders had a higher rate of adjustment of the number
 of selected students by department,expansion of division by subject, strengthening academic support(professors,
 mentoring, teaching assistants), and opportunity to participate in special activities by department.
 College of humanities had a high ratio of improvements in adjustment of the number of selected people by department,
 IT college expansion of division by subject, strengthening academic support(professors, mentoring,
 teaching assistants)., college of engineering opportunity to participate in special activities by department.
 Fourth, as a result of the morphology analysis of the requirements for multi-major operation, it was ranked first
 major, second student, third credit, fourth subject, and fifth average.
 Conclusions The results of this study are meaningful in that it provides basic data for establishing the direction
 of the multi-major curriculum. In addition, based on the results of this study, I would like to propose a way for multi-
 major customized operation in consideration of the characteristics of students.

Guidance for schoolchildren on artificial intelligence professions

The article actualizes the problem of professional orientation of schoolchildren in professions related to artificial intelligence. At the present time there is a need of society for specialists who could solve problems both on development of artificial intelligence systems, and on operation of already existing ones. The shortage of such specialists is a big problem. The aim of the article is to justify the content of career guidance activities in secondary schools in information technology (IT) professions in general and artificial intelligence (AI) in particular. Methodologically, our study was based on a theoretical analysis of the literature on career guidance for students. In the course of the study, the results of surveys of students were analyzed, documents were analyzed, and observations were made. Based on the analysis of career guidance resources and comparison with real job vacancies the specifics, specification of professions related to AI were considered. Activity-based and project-based learning methods were used in the career guidance activities. Google Klass was used as a learning platform and Google Colab as a development environment. Results. The article presents the content and the results of the career guidance master classes carried out by the authors with the schoolchildren. The results of the input and final surveys (questionnaires) of the students are analysed. Three main professions from Artificial Intelligence and their competences for the initial introduction are highlighted. Analysis of key skills of AI professions is carried out, groups of skills for career guidance are highlighted. Some essential content aspects that can be used in designing different types of career guidance sessions are suggested. In conclusion, it is concluded that the developed and tested materials contribute to professional self-determination of schoolchildren. They can be used for different career guidance activities on artificial intelligence professions.

The Framework of RBL-STEM Learning Activity: Improving Students’ Climate Change Literacy in Solving the Problem on Forecasting the Nutrition Supply of Hydroponic Plants with GNN

This paper aims to develop the learning activity framework for the RBL model integrated with the STEM approach, especially in improving the students’ climate change literacy in solving the problem on forecasting the nutritional supply of hydroponic plants using machine learning of GNN technique. It is using qualitative research which involves some bibliography study and analytical study. The findings are presented in a table containing six stages, namely stages 1-6. Each stage explains how students learn to collect data using IoT software, namely Thingspeak to collect some agriculture data, and by using Python under Google Colab platform we implement Graph Neural Networks (GNN) in RBL-STEM learning model. The main findings of this research related to RBL-STEM learning is to develop the learning activity framework in solving the problem of forecasting the nutritional needs of hydroponic plants using Thingspeak and google colab software to improve students’ climate change literacy described in stages 1-6. This research also included the development of a framework in improving the students’ climate change literacy in solving the problem on forecasting the nutritional supply of hydroponic plants using. The implication of the findings of this study is that the learning activity framework is ready to be continued in the process of developing RBL-STEM teaching materials to improve students’ climate change literacy in solving the problem on forecasting the nutritional needs of hydroponic plants machine learning of GNN technique.

Comparison of Metaheuristics in Solving the Knapsack Problem: An Experimental Analysis

Objective: Through statistical analysis using ANOVA, compare the obtained results and processing time of the metaheuristics Local Search, Tabu Search, and Genetic Algorithm programmed in Python language for application in the Knapsack Problem among the described instances. Method: The method used was modeling in order to compare randomly generated instances where the metaheuristics were programmed in Python language, inserted in Google Colaboratory, and executed in the cloud. Results and Conclusion: Analysis of Variance (ANOVA) was employed as there were three samples with paired instances to ensure conclusion validation. It was observed that, for the instances and interruption parameters of the metaheuristics used, the Genetic Algorithm generated more satisfactory results than the other metaheuristics. Research Implications: Provides relevant information about the effectiveness and performance of metaheuristic techniques, contributing to the evolution of the field of Operations Research by guiding the choice of approaches in practical applications and promoting collaboration and scientific replicability.

ANALISIS PENGARUH DIMENSI GAMBAR PADA KLASIFIKASI MOTIF BATIK DENGAN MENGGUNAKAN CONVOLUTIONAL NEURAL NETWORK

This study focuses on the classification of batik patterns using image data, specifically with three classes: kawung, parang, and pekalongan batik patterns. A total of 180 images were used for the research, divided equally among the three motifs. The dataset was collected through observations from Google. Data preprocessing involved two stages. In the first stage, the data was split into training and validation sets, with a 70% - 30% ratio, respectively. This separation allowed for evaluating the model's performance and generalization ability on unseen data. In the second stage, the Image Data Generator library was utilized to enhance data diversity and improve the model's ability to generalize patterns. Data augmentation techniques, such as rotation, shear, zoom, horizontal flip, and shift, were applied to the training data. The augmented data was then fed into a pre-defined Convolutional Neural Network (CNN) architecture. The CNN model processed the input data through convolutional layers with max pooling and ReLU activation functions. The outputs from the first convolutional layer were used as inputs for subsequent convolutional layers. The resulting feature maps were flattened and passed through fully connected layers for classification. The model's performance was assessed by evaluating accuracy measures. The CNN model achieved the highest accuracy of 98% on the training data and 100% on the validation data after 7 epochs.. The implementation was done using Python programming language and the Google Colab platform, along with required libraries. Model evaluation involved assessing the trained model's performance by inputting test data and computing evaluation metrics, including accuracy, precision, recall, and F1-score. Confusion matrix analysis provided insights into true positive, true negative, false positive, and false negative predictions. The classification report summarized the performance metrics of the model. In conclusion, the CNN method proved effective in classifying batik patterns. The size of the input images significantly influenced the accuracy of the model. A 148x148 pixel image size yielded 100% accuracy. As a suggestion, future research should consider using a larger dataset to maximize the accuracy of the classification model.

Data-driven plasma modelling: surrogate collisional radiative models of fluorocarbon plasmas from deep generative autoencoders

We have developed a deep generative model that can produce accurate optical emission spectra and colour images of an ICP plasma using only the applied coil power, electrode power, pressure and gas flows as inputs—essentially an empirical surrogate collisional radiative model. An autoencoder was trained on a dataset of 812 500 image/spectra pairs in argon, oxygen, Ar/O2, CF4/O2 and SF6/O2 plasmas in an industrial plasma etch tool, taken across the entire operating space of the tool. The autoencoder learns to encode the input data into a compressed latent representation and then decode it back to a reconstruction of the data. We learn to map the plasma tool’s inputs to the latent space and use the decoder to create a generative model. The model is very fast, taking just over 10 s to generate 10 000 measurements on a single GPU. This type of model can become a building block for a wide range of experiments and simulations. To aid this, we have released the underlying dataset of 812 500 image/spectra pairs used to train the model, the trained models and the model code for the community to accelerate the development and use of this exciting area of deep learning. Anyone can try the model, for free, on Google Colab.

Accelerating Bayesian Estimation of Solar Cell Equivalent Circuit Parameters Using JAX-Based Sampling

Equivalent circuit models that reproduce the current–voltage characteristics of solar cells are useful not only to gain physical insight into the power loss mechanisms that take place in solar cells but also for designing systems that use renewable solar energy as a power source. As mentioned in a previous paper, Bayesian estimation of equivalent circuit parameters avoids the drawbacks of nonlinear least-squares methods, such as the possibility of evaluating estimation errors. However, it requires a long computation time because the estimated values are obtained by sampling using a Markov chain Monte Carlo method. In this paper, a trial to accelerate the calculation by upgrading the Bayesian statistical package PyMC is presented. PyMC ver. 4, the next version of PyMC3 used in the previous paper, started to support the latest sampling libraries using a machine learning framework JAX, in addition to PyMC-specific methods. The acceleration effect of JAX is remarkable, achieving a calculation time of less than 1/20 times that of the case without JAX. Recommended calculation conditions were disclosed based on the results of a number of trials, and a demonstration with testable Python code on Google Colaboratory using the recommended conditions is published on GitHub.

Detection of Foreign Objects Around the Railway Line with YOLOv8

This study proposes a deep learning-based method to detect foreign objects around the railway line. It is important to make this determination with high accuracy for rail transport safety, but traditional methods are disadvantageous in terms of time and cost. In the proposed method, the RailSem19 dataset was used, and a YOLOv8-based model was designed. YOLOv8 is a prominent algorithm in the literature with its fast and accurate object detection capability. In the study, the dataset was diversified using image enhancement techniques. The training, validation, and testing stages used manually labeled data for human and car classes. The training process was carried out through Google Colab and different YOLOv8 sub-architectures were evaluated. The results showed that the YOLOv8m sub-architecture had higher mAP50 values than the other sub-architectures and showed a successful performance in the validation phase. The YOLOv8m model was able to clearly distinguish people and cars around the railway line. The YOLOv8m sub-architecture achieved a mAP50 value of 88.8%. This study presents an automated and efficient method to improve rail transport safety. The high success of the YOLOv8-based model with the RailSem19 dataset can be considered an effective solution to detect potential risks around the railway line.

Predicting photoresist sensitivity using machine learning

AbstractWe introduce a scheme for predicting photoresist sensitivity using machine learning (ML) work flow on the basis of previously reported experimental data. Different ML models, specifically Linear Regression, Kernel Ridge Regression, Gaussian Process Regressor, Random Forest Regressor, and Multilayer Perceptron Regressor, were evaluated to rapidly identify the best sensitivity prediction model. The experiment was carried out on the Google Colab platform using the Materials Simulation Toolkit for Machine Learning and Sci‐kit Learn. Different ensemble models were utilized without splitting the dataset to determine the prediction accuracy of the ML models. The hyperparameter optimization was established with a 70/30 ratio, followed by a K‐Fold cross‐validation to improve the model prediction performance. The optimized ML model showed a prediction performance of R2 = 0.83, RMSE = 10.53, and MARE = 0.68. Hence, by optimizing the hyperparameters used in the ML model, the sensitivity of the photoresist materials can be predicted with improved prediction performance.

Computationally-driven epitope identification of PEDV N-protein and its application in development of immunoassay for PEDV detection

The nucleocapsid (N) protein is a suitable candidate for early diagnosis of porcine epidemic diarrhea virus (PEDV). Here, we identified the linear B-cell epitopes of the PEDV N-protein by integrating a computational-experimental framework and constructed three-dimensional (3D) structure model of the N protein using the ColabFold program in Google Colaboratory. Furthermore, we prepared the monoclonal antibodies against the predicted epitopes and recombinant N protein, respectively, and selected pairing mAbs (named 9C4 and 3C5) to develop a double-antibody sandwich immunochromatographic test strip using CdSe/ZnS quantum dots (QDs)-labelled 9C4 and 3C5 as capture and detection antibodies, respectively. This strip can specifically detect PEDV within 10 min with a detection limit of less than 6.25 × 103 TCID50/mL. In comparison with RT-PCR for testing 90 clinical samples, the relative sensitivity and specificity of the strip were found to be 98.0% and 100%, respectively, with a concordance rate of 98.9% and a kappa value of 0.978, indicating that QDs-ICTS is a reliable method for the application of PEDV detection in clinical samples.

QuDPy: A Python-based tool for computing ultrafast non-linear optical responses

Nonlinear Optical Spectroscopy is a well-developed field with theoretical and experimental advances that have benefited multiple disciplines, including chemistry, biology, and physics. However, for the accurate interpretation of the corresponding multi-dimensional spectra, there is a need for precise quantum dynamical simulations based on model Hamiltonians. In this article, we present the initial release of our code, QuDPy (Quantum Dynamics in Python), which provides a robust numerical platform for performing quantum dynamics simulations based on model systems, including open quantum systems. A distinguishing feature of our approach is the ability to specify various high-order optical response pathways in the form of double-sided Feynman diagrams through a straightforward input syntax. This syntax outlines the time-ordering of ket-sided or bra-sided optical interactions acting on the time-evolving density matrix of the system. We utilize the quantum dynamics capabilities of QuTip to simulate the spectral response of complex systems, allowing us to compute virtually any n-th order optical response of the model system. To illustrate the utility of our approach, we provide a series of example calculations. Program summaryProgram Title:QuDPyCPC Library link to program files:https://doi.org/10.17632/5xm9pm24cz.1Developer's repository link:https://github.com/sa-shah/QuDPyLicensing provisions: MIT LicenseProgramming language: Python (v3.7)Supplementary material: Available as Google Colab Files.Example 1: https://tinyurl.com/y3j5jmmrExample 2: https://tinyurl.com/37vwntn5External packages:•QuTip (v.4.7) and dependencies i.e. Numpy, Matplotlib (https://qutip.org/)•UFSS Automatic Diagram Generator (https://github.com/peterarose/ufss)Nature of problem: Accurate quantum simulations of complex systems are required in order to understand and interpret multi-dimensional ultrafast spectroscopic signals. This code provides an open-source/multi-platform method that facilitates the generation of higher-order non-linear optical responses for an arbitrary molecular or material system given a model input Hamiltonian and bath model.Solution method: We use the double-sided Feynman diagram method [1, 2] to generate (symbolically) a set of response functions corresponding to the nth order non-linear response of the system to a series of laser pulses using the UFSS package [3] We then perform a series of accurate quantum dynamics calculations using the QuTip package [4] to generate the numerical response and spectra which correspond to specific experimental conditions.

Combination of YOLOv3 Algorithm and Blob Detection Technique in Calculating Nile Tilapia Seeds

Baby Fish counting must be counted accurately so it will not cause any loss, especially for fish seeds or fingerlings that have a small size. Generally, people still use conventional counting methods that produce low accuracy values. This research will make a Nila Baby Fish fingerlings counter program using the YOLOv3 algorithm and Blobb detection technique. The annotation data process will use LabelImg, and the dataset training will use Google COLABoratory with the Darknet framework in an online environment. Images that will predict in this program will be called and detected with an object detector. The object with a confidence score of more than 0.3 will be converted into a blob. The blob value will be forwarded to the output layer for scaling the bounding box objects. The output of this program is the predicted image, blob value, prediction time, and the number of Nila seeds. The model performance is evaluated using a confusion matrix and got a 98.87% for accuracy score.

Deteksi Otomatis Terhadap Pelanggaran Pembuang Sampah Menggunakan Metode You Only Look Once (YOLO)

Disposing of garbage is a bad thing that can spoil the view, cause bad smells, cause low to high level flooding, cause various diseases and can pollute the environment. Even though the ban on disposing of trash has been implemented, there are still many who violate it. The importance of avoiding this makes a study aimed at automatically detecting violations of waste disposal. The method used is YOLOv5, this method is an algorithm that can identify objects with high accuracy, besides that it can also carry out tracking processes in the form of bounding boxes for objects in real time. The programming language used is Google Colaboratory. The dataset used is in the form of 800 images and 2 videos. After testing the results of the research using the You only look once (YOLO) method, the best results were obtained on the parameter batch 5 epochs 5 with an accuracy of 95%. From these results it can be concluded that the use of the YOLO method is very accurate when applied to the detection process of an object.

Enhancing cryo-EM maps with 3D deep generative networks for assisting protein structure modeling.

The tertiary structures of an increasing number of biological macromolecules have been determined using cryo-electron microscopy (cryo-EM). However, there are still many cases where the resolution is not high enough to model the molecular structures with standard computational tools. If the resolution obtained is near the empirical borderline (3-4.5 Å), improvement in the map quality facilitates structure modeling. We report EM-GAN, a novel approach that modifies an input cryo-EM map to assist protein structure modeling. The method uses a 3D generative adversarial network (GAN) that has been trained on high- and low-resolution density maps to learn the density patterns, and modifies the input map to enhance its suitability for modeling. The method was tested extensively on a dataset of 65 EM maps in the resolution range of 3-6 Å and showed substantial improvements in structure modeling using popular protein structure modeling tools. https://github.com/kiharalab/EM-GAN, Google Colab: https://tinyurl.com/3ccxpttx.

SHEPHARD: a modular and extensible software architecture for analyzing and annotating large protein datasets.

The emergence of high-throughput experiments and high-resolution computational predictions has led to an explosion in the quality and volume of protein sequence annotations at proteomic scales. Unfortunately, sanity checking, integrating, and analyzing complex sequence annotations remains logistically challenging and introduces a major barrier to entry for even superficial integrative bioinformatics. To address this technical burden, we have developed SHEPHARD, a Python framework that trivializes large-scale integrative protein bioinformatics. SHEPHARD combines an object-oriented hierarchical data structure with database-like features, enabling programmatic annotation, integration, and analysis of complex datatypes. Importantly SHEPHARD is easy to use and enables a Pythonic interrogation of largescale protein datasets with millions of unique annotations. We use SHEPHARD to examine three orthogonal proteome-wide questions relating protein sequence to molecular function, illustrating its ability to uncover novel biology. We provided SHEPHARD as both a stand-alone software package (https://github.com/holehouse-lab/shephard), and as a Google Colab notebook with a collection of precomputed proteome-wide annotations (https://github.com/holehouse-lab/shephard-colab). Supplementary data are available at Bioinformatics online.