English 101

Final Project

English Reflective Portfolio

Gaurvendra Pratap Singh Pundhir

Introduction

This portfolio is a reflection of my journey as a student pursuing a Bachelor of Science in Computer Science at the University of Arizona. My name is Gaurvendra Pratap Singh Pundhir, and I aspire to become a software engineer and developer. In addition to my academic pursuits, I am also working on my tech start-up. Writing was not something that I was initially interested in, but with new upgrades and necessities, I have come to realize that writing is an essential skill that is used in every aspect of life, whether it’s writing an email or a memo for a start-up company pitch. This portfolio showcases my growth in specific skills that I have developed throughout the course. Each artifact represents a milestone in my journey, and I hope that this portfolio serves as a testament to my hard work and dedication. Throughout the course, I have developed specific skills that are essential for my future career as a software engineer and developer. These skills include rhetorical awareness, critical thinking and composing, conventions, revision, and reflective practice. I have learned how to tailor communication for diverse audiences, synthesize information critically, leverage writing conventions strategically, refine my writing iteratively, and transform challenges into learning opportunities. The artifacts in this portfolio represent not just final products but also milestones in my development of these skills. I understand that writing is an essential skill that is used in every aspect of life, and I am committed to honing my craft. I hope that this portfolio serves as a testament to my hard work and dedication, and I look forward to continuing to grow and develop as a writer and a software engineer.

Introduction

This reflective portfolio is a testament to my transformative journey throughout the course. Each artifact represents a milestone in the development of specific skills, including:

1. Rhetorical Awareness: I have developed the ability to tailor communication for diverse audiences, making complex technical concepts accessible without compromising depth.
2. Critical Thinking and Composing: I have a heightened capacity to synthesize information critically, contextualizing complex ideas and understanding the broader societal implications of technology.
3. Conventions: I leverage writing conventions strategically for enhanced communication impact, moving from necessity to intentional choice. For instance, the exploration into motion graphics in The Boat represents a departure from traditional video genres, adding a layer of engagement and realism.

4. Revision: I have refined my skill in iterative refinement, recognizing that revision is not just about correcting errors but strategically enhancing overall writing quality.
5. Reflective Practice: I have an evolved metacognitive process, transforming challenges into learning opportunities and celebrating successful research strategies.

Artifacts

1. Rhetorical Awareness - Project 2 Final:

I found Project 2 Final to be a turning point in my rhetorical awareness. The complexity of AI required me to communicate in a nuanced way. Throughout the project, I honed my ability to tailor the narrative for diverse audiences. Navigating the intricacies of Transformers and attention mechanisms demanded clarity without compromising depth. By employing clear language, effective illustrations, and strategic structuring, I successfully conveyed the nuances of large language models. This project elevated my skill in adapting rhetorical techniques to meet the demands of a multifaceted audience.

2. Critical Thinking and Composing - Project 1 Final:

Project 1 Final, which addressed the profound impact of AI in education, showcased substantial growth in my critical thinking and composing skills. I synthesized information from diverse sources to weave a coherent narrative that resonates with readers. The intentional inclusion of real-world examples illustrated my ability to contextualize complex ideas. This project deepened my understanding of AI’s societal implications, fostering critical thinking not only about the technology but also its broader impact. My composing skills matured, resulting in an essay that seamlessly blended technical insights with real-world relevance.

Artifacts

3. Conventions - The Boat Notes :

The Boat Notes underscored my commitment to writing conventions. I paid meticulous attention to detail, allowing me to leverage conventions strategically for enhanced communication impact. Font choices, formatting, and language style were transformed from mere elements into tools strategically employed. This project improved my adherence to conventions and instilled an awareness of how conventions contribute to overall communication effectiveness. The inclusion of motion graphics in The Boat represents a departure from traditional video genres, adding engagement and realism. This exploration into motion graphics showcased my awareness of multimedia possibilities and a willingness to embrace innovative presentation methods.

4. Revision - Project 1 and 2 Drafts:

The comparative analysis of Project 1 and 2 drafts revealed pronounced improvement in my revision skills. Multiple iterations honed my ability to refine ideas, structures, and language choices strategically. Tangible evidence showcased my commitment to continuous improvement. I learned that revision is not just about correcting errors but a strategic process of refining and enhancing overall writing quality. The iterative nature of this project emphasized the evolution of my writing, reflecting a more mature and intentional approach to the craft.

Artifacts

5. Reflection - Project 1 & 2 Research Reflection:

Reflecting on the research phases of Project 1 and 2

, I realized that I had undergone substantial personal and intellectual growth.
My metacognitive processes evolved, and I became adept at self-assessment.
Challenges became learning opportunities, and successful research strategies were identified and celebrated.

This reflection underscored not only the complexities of synthesizing information but also the development of resilience and adaptability.

It encapsulated a journey of personal and intellectual growth, showcasing a refined ability to navigate the intricacies of research and writing.

Reflection on Research for Project 2:

When I embarked on the research journey for Project 2, I delved into the intricate world of large language models and their decipherability. The overarching question, “Can we decode large language models?” prompted a comprehensive exploration, leading to a series of thought-provoking questions and insights.

1. Defining Large Language Models:

o I explored the fundamental definition of a language model in machine learning. o I investigated how these models predict and generate plausible language. o I questioned whether autocomplete can be considered a quintessential example of a language model.

2. Understanding Models in General:

o I delved into how we define a “model” in the context of machine learning. o I examined the role that parameters and structure play in the functionality of a model. o I explored the ways in which models process input data and generate output.

3. Evolution with Transformers:

o I researched the pivotal role that Transformers played in the evolution of language modelling. o I investigated how Transformers addressed memory issues encountered in earlier models.

4. The Significance of Attention:

• The attention mechanism within a neural network is a technique that allows the network to focus on specific parts of the input data that are most relevant to the task at hand. It prioritizes and emphasizes relevant information, acting as a spotlight to enhance overall model performance. • Attention compresses the information needed for predicting the next token/word by selectively focusing on important input elements, improving prediction accuracy and computational efficiency. • Attention is crucial in enhancing the efficiency of language models because it allows the model to focus on the most relevant parts of the input data, reducing the computational burden and improving the accuracy of predictions.

5. Weights and Inference:

• Weights play a crucial role in the functioning of a machine learning model. They control the signal (or the strength of the connection) between two neurons, deciding how much influence the input will have on the output. • Training is integral to determining a model’s ideal weights. During training, the model learns to adjust its weights to minimize the difference between its predicted output and the actual output. • The process of inference contributes to making predictions in machine learning by applying the trained model to new data. During inference, the model uses its learned weights to generate predictions based on the input data.

6. Neural Networks and Transformers:

• A neural network is a type of machine learning model that is inspired by the structure and function of the human brain. A deep neural network is a neural network with multiple hidden layers. • Transformer models are a type of neural network architecture that have revolutionized natural language processing. They are characterized by their use of self-attention mechanisms, which allow them to selectively focus on different parts of the input data to generate more accurate predictions. • Transformer models apply attention or self-attention to understand sequential data by selectively focusing on important input elements, improving prediction accuracy and computational efficiency.

7. Mechanistic Interpretability:

• Mechanistic interpretability is a subfield of AI interpretability that focuses on reverse-engineering neural networks. The goal is to reverse engineer the parameters of a trained neural network and try to reverse engineer what algorithms and internal cognition the model is actually doing. • Mechanistic interpretability focuses on reverse-engineering neural networks to uncover the learned algorithms within these networks. This subfield aims to provide a quantitative, interpretable description of how a neural network solves a particular task. • Mechanistic interpretability aims to uncover the learned algorithms within neural networks by reverse-engineering the parameters of a trained neural network. This subfield focuses on understanding the internal cognition of the model and how it solves a particular task.

8. The Essence of Reverse Engineering:

• Reverse engineering entails the process of taking apart a system to see how it works.

When applied to neural networks, reverse engineering involves trying to reverse engineer the parameters of a trained neural network to understand the internal cognition of the model.


• The property of decomposability is crucial in the reverse engineering process because it allows us to break down the model into smaller, more manageable parts. This facilitates reasoning about a model without fitting its entirety in our head.


• Decomposability facilitates reasoning about a model without fitting its entirety in our head by breaking down the model into smaller, more manageable parts. This allows us to understand the internal cognition of the model and how it solves a particular task.

Conclusion

Based on my project, I have developed a range of skills that will be useful in my future endeavours. I have demonstrated the ability to communicate complex technical concepts to diverse audiences without compromising depth, which is a valuable skill in today’s world. Additionally, I have shown a heightened capacity for critical thinking and composing, which will enable me to contextualize complex ideas and understand the broader societal implications of technology. I have also leveraged writing conventions strategically to enhance communication impact, as evidenced by my exploration into motion graphics in The Boat. Furthermore, I have refined my skill in iterative refinement, recognizing that revision is not just about correcting errors but strategically enhancing overall writing quality. Finally, I have an evolved metacognitive process, transforming challenges into learning opportunities and celebrating successful research strategies. I am proud of my achievements and excited to apply these skills in my future work!