| 奥野 智也 | M, 2回目発表 | 計算システムズ生物学 | 金谷 重彦 | 宮尾 知幸 | 小野 直亮 | MD.Altaf-Ul-Amin |
|
title: Feature Learning for Enzyme Based on Amino Acid Sequence and Enzymatic Reaction
The identification of substrates of enzyme is critical for leveraging enzyme in industrial applications. To investigate the potential combinatorial space of enzyme and molecule, data-driven approaches have been applied in the literature. This study aims to improve the performance of machine learning models on the enzyme substrate pair activity prediction task by utilizing a diverse range of enzyme data available in public databases. Masked language modeling on enzymatic reaction data is employed to adapt a pre-trained amino acid sequence language model as an amino acid sequence feature extractor. The performance of this amino acid sequence feature extractor is evaluated through potential enzyme and substrate pair activity binary classification on family wide screening data. Although the current prototype model does not outperform existing methods, it demonstrates an improvement over baseline approaches. Further research and model refinement are necessary to achieve better results. language of the presentation: Japanese | ||||||
| 土橋 有理 | M, 2回目発表 | ユビキタスコンピューティングシステム | 安本 慶一 | 金谷 重彦 | 諏訪 博彦 | 松田 裕貴 |
|
title:A Behavior Change System to Promote Exercise and Human Interaction among the Elderly
abstract:Due to the declining birthrate and aging population, as well as changes in family structure, the elderly are going out less frequently, resulting in social isolation and frailness becoming more serious. In this study, we developed a behavior change system utilizing Bluetooth Low Energy (BLE) signals to promote outings and social interaction among the elderly. The system combines tags that record BLE signals with digital signage that displays information to visualize and share the outing activities of the elderly.The system calculates the distance traveled and places visited by individuals based on the data collected by the tags, and grows flowers on the digital signage accordingly. The display of flowers that have grown in proportion to the distance is intended to provide an incentive for the elderly to go out and engage in activities, and also to promote interaction among residents through signage installed in the common areas of housing complexes.We proposed a method to estimate the location of the elderly based on the reception strength (RSSI) of the BLE signal, and confirmed in preliminary experiments that approximate location estimation is possible. language of the presentation: Japanese | ||||||
| 中川 翔太 | D, 中間発表 | 計算システムズ生物学 | 金谷 重彦 | 松本 健一 | 小野 直亮 | MD.Altaf-Ul-Amin |
|
title: Quantitative evaluation model of variable diagnosis for chest X-ray images using deep learning
abstract: My research aims to quantify uncertainty in diagnostic evaluation using deep learning models on chest X-ray images. Specifically, it focuses on conditions with high diagnostic variability among physicians, such as pleural thickening and scoliosis, using a dataset annotated with diagnoses from multiple doctors. Seven different deep learning models (six pre-trained models and one autoencoder-based model) were employed to assess the correlation between the model output values and the expected values of diagnostic probabilities, modeled as binary processes based on multiple physicians' diagnoses. The results demonstrated high correlations (0.89 to 0.97), indicating that deep learning models can effectively quantify uncertainty in diagnostic evaluations. language of the presentation: Japanese | ||||||
| INDIRA FEBRIYANTI | M, 2回目発表 | ソフトウェア工学 | Raula Gaikovina Kula | 安本 慶一 | 松本 健一 | 嶋利 一真 |
|
title: An Empirical Study On The Relationship Between Proficient Code and Maintainability in Python Libraries
abstract: Python is very popular because it can be used by a wider audience of developers, data scientists, machine learning experts, and so on. Like other programming languages, there are beginner to advanced levels of writing Python code. However, like all software, code constantly needs to be maintained as bugs and the need for new features emerge. Although the Zen of Python states that “Simple is better than complex,” I hypothesize that more developers tried to write proficient code might be harder for the developer to maintain, but does proficient code necessarily ensure maintainability? To study this relationship between the understanding of code proficiency and code maintainability, I present an exploratory study into the complexity of Python code on three Python libraries. Specifically, the study investigates the risk level of proficient code inside a file. As a starting point, I mined and collected code proficiency from three PyPI libraries totaling 3,003 files. I identified several instances of highly proficient code that were also high risk. My early examples revealed that most code-proficient development presented a low maintainability risk, yet there are some cases where proficient code is also risky to maintain. This study should help developers identify scenarios where and when using proficient code might be detrimental to future code maintenance activities. language of the presentation: English | ||||||