50+ days of Machine Learning

 Hello :) Today is Day 61! A quick summary of today: Covered Lecture 7 : machine translation, seq2seq, attention from Stanford CS224N Tried to make a conditional DCGAN to generate MNIST numbers ( colab ) ( kaggle ) I will first cover the GAN story (then will share my notes from the lecture) So... while watching and taking notes today, I started thinking, what if I can use my notes as data to a model and afterwards, when I want, I can give it raw string text and it will output text in the format of my notes (with my handwriting). Well I started looking around and actually the first model architecture that came to my mind was the GAN (specifically conditional GAN) - I remembered there was a GAN architecture that alongside the pictures, we can give it the labels, and then on-demand generate. In retrospect, there are of course others, but I decided to go with GAN.  For maybe 2 hours I busted my head trying to make a simple model with the EMNIST dataset (english characters), and I ...

 Hello :) Today is Day 60! A quick summary of today: Covered lecture 5 and 6 from CS224N Attempted  assignment 3 - Neural Transition-Based Dependency Parsing So far, including today, the lectures have been just phenomenal. Based on my previous NLP knowledge and now with these lectures that dive deeper into how modern language models came to be - I am establishing a solid foundation. My notes from the lectures are below: Lecture 5: Language models and RNNs Lecture 6: Simple and LSTM RNNs Tomorrow's turn is Lecture 7: Machine Translation, Attention, Subword Models and Lecture 8: Transformers depending on lecture 7's length and readings. That is all for today! See you tomorrow :)

 Hello :) Today is Day 59! A quick summary of today: From the  Stanford CS224N course , I covered: Lecture 3: Backprop and Neural Networks Lecture 4: Dependency Parsing And below are my notes for both.  Lecture 3: Backprop After Andrej Karpathy, backprop and I have a friend-friend relationship, so this felt like a nice overview over backprop in NNs. Lecture 4:  Dependency Parsing This one felt like a linguistics lesson, learning about how people interpret language and how such logic transferred to computers. Tomorrow is RNN's turn. Really exciting! That is all for today! See you tomorrow :)

 Hello :) Today is Day 58! A quick summary of today: covered Lecture 2 of Stanford's NLP with DL did assignment 1  on google colab which covered some exercises on count-based and prediction-based methods Read 6 papers about word embeddings and wrote down some basic summaries  GloVe: Global Vectors for Word Representation Improving Distributional Similarity with Lessons Learned from Word Embeddings Evaluation methods for unsupervised word embeddings A Latent Variable Model Approach to PMI-based Word Embeddings Linear Algebraic Structure of Word Senses, with Applications to Polysemy On the Dimensionality of Word Embedding Firstly, I will share my notes from the lecture: Next, are the short summaries of each paper. These were definitely interesting papers, I feel like I am learning the history of something big, and after a few lectures I will be in the present haha. These papers, that preceeded the infamous transformer, and that laid grounds for modern embeddings were v...