Teachable machine Research Paper Rough Draft
Corinne Hrnicek
Phoenix College Stem Train Program
Chapman
31 October 2023
Abstract
In summary, this project is training machine learning/teachable machines to identify bacteria. This was done by culturing a few different kinds of bacteria, taking pictures of said bacteria, and uploading the images to a teachable machine. Once this was complete, the software was tested by seeing if it can identify the bacteria using the information it was given. Results are that it is possible to train the computer to consistently recognize bacteria from pictures after it has been trained using other pictures of the same bacteria. This can have positive ramifications in the world of science and medicine because with teachable machines, bacteria can be accurately identified faster and easier than before.
Project Background
Research question: Can teachable machine learning be trained to identify bacteria accurately? Hypothesis: If we train teachable machine learning, it will accurately identify bacteria.
What is currently known about the topic is that there is a “well-known system, Mycin, which can diagnose bacterial infections. The input to Mycin is the set of symptoms experienced by a patient. Based on a set of rules triggered by each symptom, Mycin suggests the type of bacteria that could cause the problem.” (Lerner). “The researchers plan to use their platform to study more bacteria and media types, using the information to build a training data library of various bacterial types in additional media to reduce the collection and detection times for new samples.” (New)
It is important to expand on this current knowledge because “Expert systems cannot pass the Turing test because of their limited domain of expertise.”(Lerner) We can do more research and improve accuracy. Another reason this research is important is that “This approach eliminates the need for time-consuming culture-based colony isolation and resource-intensive molecular approaches for bacterial identification.”(Ma). It is important to work on this specifically because “ relatively few studies have been conducted to optimize BPA removal using microalgae-bacteria consortia and artificial intelligence to predict degradation rate, and even fewer research have been conducted on the kinetic models in the degradation of BPA. (Fu)
This topic was selected because it is significant in medicine, biology, and technology. Learning more about this topic and researching to identify bacteria faster is essential. An example of how teachable machine learning can identify bacteria in a medical setting is stated in the article. “Gram-negative bacteria (ESBL-GNB) accounting for individual- and group-level confounding using machine-learning methods. Patients hospitalized between September 2010 and June 2013 at six medical and surgical wards in Italy, Serbia and Romania were screened for ESBL-GNB”(Tacconelli). That is why it is important to research this topic.
The research will be conducted by observing bacteria, taking pictures, and experimenting with teachable machine learning. “In the last decade, NIR spectroscopy has been applied to identify bacteria, fungi, and viruses, combined with Machine Learning algorithms and Multivariate Analysis techniques”(Farias,). In a similar way, the bacteria will be observed when experimenting with a teachable machine.
“The clustering algorithms…applied to the dataset) and their performance was evaluated by means of accuracy”(Duran) I anticipate finding that teachable machine learning can be trained to Identify bacteria. “Ranking and selection of probiotic potential bacteria for harnessing as antibacterial agents in plant tissue cultures were performed using supervised machine learning models.”(Sadeghi). Another source also supports the hypothesis that “a machine learning-based method can perform fast estimation of the concentration of indoor airborne culturable bacteria. “(Liu)
Variable table
Name
I/D/C
Symbol
Units
Description
Teachable Machine Learning program
I
I
?
Algorithm to test accuracy of machine learning
Accuracy
D
A
%
Measure of correct prediction
Data Set
C
D
?
Collection of Date
Data
Procedure:
Spread bacteria in a petri dish and culture bacteria in incubator (37 degrees celsius)
After 24 hours take pictures of microscope view of bacteria
Upload pictures to teachable machine
Test a picture of the bacteria to see that teachable machine identifies it
Analysis
Given the evidence that one hundred percent of the time bacteria were correctly identified, that means the hypothesis was correct
Claim
Yes, Teachable machines can accurately identify bacteria when trained.
Evidence
Reasoning
Merit
This research is important because it could make it easier and faster to accurately identify bacteria which is useful in everyday life in treating, killing bacteria, and stopping the spread of bacteria by identifying it sooner.
References
Durán, C., Ciucci, S., Palladini, A., Ijaz, U. Z., Zippo, A. G., Sterbini, F. P., Masucci, L., Cammarota, G.,
Ianiro, G., Spuul, P., Schroeder, M., Grill, S. W., Parsons, B. N., Pritchard, D. M., Posteraro, B., Sanguinetti, M., Gasbarrini, G., Gasbarrini, A., & Cannistraci, C. V. (2021). Nonlinear machine learning pattern recognition and bacteria-metabolite multilayer network analysis of perturbed gastric microbiome. Nature Communications, 12(1), 1926–1926. https://doi.org/10.1038/s41467-021-22135-x
Farias, L. R., Panero, J. D., Riss, J. S., Correa, A. P., Vital, M. J., & Panero, F. D. (2023). Rapid and
Green Classification Method of Bacteria Using Machine Learning and NIR Spectroscopy. Sensors (Basel, Switzerland), 23(17), 7336. https://doi.org/10.3390/s23177336
Fu, Wenxian, et al. “Enhanced degradation of bisphenol A: Influence of optimization of removal,
kinetic model studies, application of machine learning and microalgae-bacteria consortia.” The Science of the Total Environment, vol. 858, 2023, p. 159876, https://doi.org/10.1016/j.scitotenv.2022.159876.
Lerner, K. L. (2022). Artificial Intelligence. In Gale Science Online Collection. Gale.
https://link-gale-com.ezproxy.pc.maricopa.edu/apps/doc/XJYXMI242056316/SCIC?u=mcc_phoe&sid=summon&xid=2e8ed34d
Liu, Z., Li, H., & Cao, G. (2017). Quick Estimation Model for Indoor Airborne Culturable Bacteria
Concentration: An Application of Machine Learning. International Journal of Environmental Research and Public Health, 14(8), 857. https://doi.org/10.3390/ijerph14080857
Ma, L., Yi, J., Wisuthiphaet, N., Earles, M., & Nitin, N. (2023). Accelerating the
Detection of Bacteria in Food Using Artificial Intelligence and Optical Imaging. Applied and environmental microbiology, 89(1), e0182822. https://doi.org/10.1128/aem.01828-22
"New Machine Learning Technique Identifies Different Bacteria in Seconds." Clinical Lab
Products, vol. 52, no. 2, Mar.-Apr. 2022, p. 6. Gale Health and Wellness, link.gale.com/apps/doc/A702628059/HWRC?u=mcc_mesa&sid=summon&xid=d746a4f3. Accessed 4 Oct. 2023.
Sadeghi, M., Panahi, B., Mazlumi, A., Hejazi, M. A., Komi, D. E., & Nami, Y. (2022). Screening of
potential probiotic lactic acid bacteria with antimicrobial properties and selection of superior bacteria for application as biocontrol using machine learning models. Food Science & Technology, 162, 113471. https://doi.org/10.1016/j.lwt.2022.113471
Söylemez, Ü. G., Yousef, M., Kesmen, Z., Büyükkiraz, M. E., & Bakir-Gungor, B. (2022). Prediction of
Linear Cationic Antimicrobial Peptides Active against Gram-Negative and Gram-Positive Bacteria Based on Machine Learning Models. Applied Sciences, 12(7), 3631. https://doi.org/10.3390/app12073631
Tacconelli, E., Górska, A., De Angelis, G., Lammens, C., Restuccia, G., Schrenzel, J., Huson, D. H.,
Carević, B., Preoţescu, L., Carmeli, Y., Kazma, M., Spanu, T., Carrara, E., Malhotra-Kumar, S., & Gladstone, B. P. (2020). Estimating the association between antibiotic exposure and colonization with extended-spectrum β-lactamase-producing Gram-negative bacteria using machine learning methods: a multicentre, prospective cohort study. Clinical Microbiology and Infection, 26(1), 87–94. https://doi.org/10.1016/j.cmi.2019.05.013
