Intelligent assistants like Alexa and Siri, image searches that found the topic of a photo, and self-driving cars – these intelligent systems used machine learning to develop their expertise. In this module, I learned about various machine learning techniques but concentrated on deep neural learning. I learned about the underlying theory and got hands-on experience creating, training, evaluating, and using machine learning systems. I also examined how these technologies were used and misused and what that meant for our societies and communities.

 

Learning Outcomes

  • Understanding Machine Learning's Social Impacts: I gained an insight into the broader issues surrounding machine learning, including its utilization and societal effects.
  • Deep Neural Learning Mastery: I acquired skills in creating, training, and evaluating neural networks for tasks like handwriting recognition, and understood their limitations in scaling to larger problems.
  • Proficiency with Convolutional Neural Networks (CNNs): I learned how CNNs identify and combine features in inputs to achieve image classification, enhancing my ability to work with image-based data.
  • Competence in Recurrent Neural Networks (RNNs): I developed the ability to use RNNs for time-dependent data analysis, such as language processing, leveraging their capability to retain information for contextual interpretation.
  • Autoencoder Applications: I explored autoencoders for data compression, cleaning, and missing data replacement, understanding their role in generating deepfakes and the implications for digital trust.
  • Comparative Analysis of Machine Learning Models: I compared deep learning with alternative machine learning approaches, assessing their strengths, weaknesses, and the biases that can arise from data preparation.
  • Critical Evaluation of Machine Learning Systems: I learned to critically review the technical aspects of machine learning systems, their applications, and the potential impacts on individuals and society.

Feedback on my assingments

  • Assignment 1: Feedback emphasized the importance of using appropriate visualizations, such as scatter plots, and highlighted areas for code improvement and better utilization of provided functions. Commendation was given for a good understanding of datasets and training processes.
  • Assignment 2: Noted avoidable mistakes and urged clarity in understanding and interpreting model accuracy. Positive recognition was given for the comprehension of NLP ethics and the application of concepts from academic papers.
  • Assignment 3: Encouraged more detailed explanation of model parameters, improvement in referencing, and clearer data description. Feedback also pointed out inaccuracies in the confusion matrix and suggested enhancements in methodology and data collection for decision-making relevance.

Reflections

  • The feedback has been instrumental in highlighting my strengths, such as engaging writing and a good grasp of data science concepts. However, it also underscored areas needing attention, such as precision in model evaluation and the interpretation of results. I appreciate the constructive critique on visualization choices and coding practices, recognizing the importance of continuous learning and adherence to best practices in data science.

Professional Skills Matrix learnt and Action Plan

 

Skills Gained or Enhanced:

  • Assignment 1: Data Visualization, Code Optimization, Understanding of Machine Learning Models
  • Assignment 2: Model Evaluation, Interpretation of Results, Ethical Considerations in NLP
  • Assignment 3: Technical Writing, Model Parameterization, Research Methodology

Action Plan:

  • Assignment 1: Dive deeper into data visualization techniques, ensuring the correct graph types for various data interpretations. Revisit and practice coding best practices to avoid redundancy and enhance code efficiency.
  • Assignment 1: Focus on rigorous model evaluation techniques, including comprehensive understanding and application of performance metrics. Engage with recent literature on ethical AI to sharpen ethical reasoning in NLP projects.
  • Assignment 1: Enhance technical writing to be more precise and formal, particularly in explaining model choices and parameters. Commit to a detailed study of hyper-parameters' impact on model performance. Improve research skills by engaging with more academic papers, focusing on critical reading and application of research findings.