-
Sequential decision-making is a fundamental challenge in machine learning and AI. From planning your next vacation itinerary to training a robot to navigate a warehouse, we often face tasks where a series of actions must be taken to achieve a goal, often with delayed feedback or sparse rewards. [Read More]
-
From 1+1 in Assembly to LLMs: The Evolution of Computing Abstraction
Tracing the Layers from Machine Code to Natural Language Interfaces
By Taewoon KimComputing has come a long way since the early days of punch cards and assembly language. With each new generation of programming paradigms, we’ve added layers of abstraction that make it easier for humans to interact with machines. In this post, we’ll explore how a simple operation like 1 +... [Read More] -
The Problems with p-values
Why Frequentist Significance Testing Falls Short
By Taewoon KimStatistical significance testing, specifically the use of p-values, has been the cornerstone of hypothesis testing for decades. However, this frequentist approach has critical flaws that can lead to misleading interpretations and false confidence in research results. In this post, we will break down why p-values often fall short and discuss... [Read More] -
Can the Transformer be viewed as a special case of a Graph Neural Network (GNN)?
A natural language text can be seen as a knowledge graph
By Taewoon KimIn recent years, Transformers have dominated the field of natural language processing (NLP), while Graph Neural Networks (GNNs) have proven essential for tasks involving graph-structured data. Interestingly, the Transformer can be seen as a special case of GNNs, particularly as an attention-based GNN. This connection emerges when we treat natural... [Read More] -
Is supervised learning a special type of reinforcement learning?
The reason why reinforcement learning is such a hard problem
By Taewoon KimSupervised Learning Objective: Maximum Likelihood [Read More] -
Playing around with Hugging Face Llama 3.1 locally
4bit quantization is amazing
By Taewoon KimLlama 3.1 was released by Meta a month ago, and you can easily access it via Hugging Face. I used their Transformers quite extensively a few years ago, and haven’t used it for some years. It was already an amazing library back then but now it has become even easier.... [Read More] -
Training a GCN-based edge classifier
Exploring the challenges and strategies for effective edge classification with graph neural networks (GNNs)
By Taewoon KimIn this post, I’ll walk you through the process of training a Graph Convolutional Network (GCN) for edge classification, a common task in graph-based machine learning applications. Edge classification involves predicting the type of relationship (or edge) between two nodes in a graph, which is particularly useful in areas like... [Read More] -
A machine with human-like memory systems
Can machines think like us?
By Taewoon KimThe project “A Machine With Human-Like Memory Systems” is the core of my PhD work. It was heavily inspired by the cognitive science theories, such as the ones from Endel Tulving. It’s about developing agents equipped with human-like external memory systems, modeled using knowledge graphs. These agents are designed to... [Read More] -
Using ChatGPT is just so great
This really can do a lot of things, although it's still biased.
By Taewoon KimGenerating images with text [Read More] -
My new website
My boring old website had to be gone.
By Taewoon KimI thought that I can make a better website than the one I had. So I made this one! I’ll be posting my projects and other stuff here. I hope you like it! :) I was surprised that I could make this website in just a few hours. I used... [Read More] -
Bloom: A 176b-parameter open-access multilingual language model
Unlocking the power of open-source language models
By Taewoon KimThis paper was a result of the Hugging Face BigScience research workshop that I participated in 2021. The paper can be found at https://arxiv.org/abs/2211.05100. [Read More] -
Room classifier
A simple room classifier made using EfficientNet and PyTorch Lightning
By Taewoon KimCheck out https://github.com/tae898/room-classification -
Promptsource: An integrated development environment and repository for natural language prompts
Revolutionizing NLP with a collaborative hub for crafting and sharing prompts
By Taewoon KimThis paper was a result of the Hugging Face BigScience research workshop that I participated in 2021. The paper can be found at https://arxiv.org/abs/2202.01279. [Read More] -
IGLU: interactive grounded language understanding in a collaborative environment
We won the NeurIPS 2021 competition
By Taewoon KimThank you, Hybrid Intelligence team! Check out https://proceedings.mlr.press/v176/kiseleva22a/kiseleva22a.pdf [Read More] -
Multitask prompted training enables zero-shot task generalization
Achieving superior zero-shot generalization through explicit multitask prompting
By Taewoon KimThis paper was a result of the Hugging Face BigScience research workshop that I participated in 2021. The paper can be found at https://arxiv.org/abs/2110.08207. [Read More] -
EmoBERTa: Speaker-aware emotion recognition in conversation with RoBERTa
Achieving state-of-the-rrt emotion recognition in conversations with a simple RoBERTa-based approach
By Taewoon KimThis is a paper that I wrote together with Piek Vossen. We were able to achieve SOTA back then by simply training a RoBERTa (a variant of BERT) with speaker tokens! Check out https://arxiv.org/abs/2108.12009 [Read More] -
Generalizing MLPs with dropouts, batch normalization, and skip connections
A simple age-gender classification model using Arcface embeddings and MLPs
By Taewoon KimI trained a simple age-gender classification model using ArcFace embeddings and MLPs. Check out the paper https://arxiv.org/abs/2108.08186 and the code https://github.com/tae898/age-gender/ [Read More]