Welcome back to my Outreachy blog series, where I share my progress and experiences during my Outreachy internship.
I am Nikhila K S an undergraduate student from IGDTUW India (Socials Link Tree). If you want to know more about Outreachy check out my previous blogs🌸.
Now back to week 2😄
My co-intern Sumaya and I embarked on an exciting adventure into the world of machine learning. Join us as we explore different tools, datasets, and learning opportunities that have shaped our understanding and enthusiasm for this fascinating field.
⚡Exploring Tools: Google Teachable Machine, Edge Impulse, and Google Colab
During this week, Sumaya, who had prior experience in machine learning, and I, an absolute beginner, ventured into the vast realm of machine learning. We began our exploration by investigating various tools like Google Teachable Machine, Edge Impulse, and Google Colab, all of which held promise for developing our machine-learning model. After careful consideration, we concluded that Google Colab provided the flexibility and control we needed for our project. Its extensive library and resources made it the ideal platform for building our machine-learning model.
⚡Discovering Valuable Datasets on Kaggle
To effectively train our machine learning model, we set out to find suitable datasets. Our journey led us to Kaggle, a renowned platform for data science enthusiasts. Here, we discovered a plethora of valuable datasets that aligned perfectly with our project requirements. These datasets would serve as the foundation for training and evaluating our model accurately.
Some of our References -
⚡Choosing the Optimal Cloud Storage Solution: Google Drive
With an abundance of data at our disposal, we needed a reliable cloud storage solution. We explored various options, such as Google Drive, Mega Account, and Firestore, to determine the most suitable choice. After careful deliberation, we unanimously decided to move forward with Google Drive. Its seamless integration with other Google services and user-friendly interface made it the ideal platform for organizing and managing our project files.
⚡Gaining Insights from Similar Projects-
To enhance our understanding of machine learning concepts and explore different approaches, we dedicated time to studying repositories and documentation related to similar projects. This deep dive allowed us to glean valuable insights and identify potential strategies to successfully complete our project. By building upon the knowledge and experiences shared by others, we could shape our own unique approach.
⚡Research Findings
How to train the heart or lung's sound data to detect anomalies.
Digital Stethoscope AI App
This project was shared by mentor @ruqaiya
Uses ML libraries from Google Tensorflow to train sound data. It looks out for anomalies in sound leveraging the features of the sound. After which it then classifies the sounds into various types of heart/lung conditions(respiratory tract infection e.g. COVID-19 or Pneumonia). Among the Classification labels is Normal.Bird Sound Identifier on Native Android Project
This project identified bird sound using the frequency of the sound. It deployed the trained data in tflite data type to Native Android for testing. It also has a much better version of where it began to look out for the types of birds-🔗[This video combines the bird sound identifier with a further analysis of bird type, it has a corresponding video made by Tensorflow on how to achieve it].Sound classification with YAMNet
This is a Tensorflow project for Classifying sound data. A couple of projects we looked into used this implementation. It follows the most basic approach to Classifying sound data leveraging the TensorFlow library.
⚡Mentor Meeting and Positive Reviews🚀
Over the weekend, we had the opportunity to present our progress during a mentor meeting. Through a detailed presentation, we showcased our achievements and received positive reviews from our mentors. This feedback served as a motivation booster, fueling our determination to excel in our project and exceed expectations.
⚡Embracing the Learning Journey in Machine Learning
As a beginner in machine learning, I found this week's exploration to be both exciting and enjoyable. Although learning machine learning in a week is a formidable task, I firmly believe that continuous learning while actively working on our project tasks will significantly enhance my understanding and expertise in this field. To deepen my knowledge, I delved into Sound classification with YAMNet and compiled detailed notes, solidifying my grasp of the topic. (File link)
⚡A new habit🤗...
I embraced a new habit of documenting my learning journey on GitHub, recognizing its advantages as a centralized hub. This practice not only enables me to quickly revisit and reinforce my understanding of concepts but also eliminates the need for excessive computer storage. GitHub has now become my reliable one-stop destination to reflect on and review my progress whenever I desire. If you are starting with a new technology I would suggest you to do the same🚀.
Wanna check out my repo? here you go, I will keep updating the repo as I learn - Unravel_ML
Week 2 of my Outreachy internship immersed me in the captivating world of machine learning, leaving an indelible mark on my journey⚡. As I conclude this blog, I carry the spirit of continuous improvement, driving me to surpass achievements each week.
With an insatiable thirst for knowledge, I eagerly embrace the upcoming weeks, equipped with newfound expertise and an unwavering determination to thrive in my Outreachy project. Stay connected as I navigate this exhilarating path, where I will unravel the endless possibilities of machine learning, making a lasting impact through our dedication, passion, and unyielding pursuit of success.
Have a nice day🌸
Top comments (0)