bona's log

Before.. RNN LSTM : slow to train Can we parallelize sequential data? Transformers Input sequence can be transmitted parallel No concept of time step Pass all the words simultaneously and determine the word embedding simultaneously (RNN passes input word one after another) Input Embedding In embedding space, close-meaning words locate close to each other There're already pretrained embedding spa..

Docker uses Linux network namespaces to isolate the container networks from the host network. When a docker container is created, docker creates a separate network namespace and puts the container into it. Then, Docker connects the new container network to Linux bridge docker0 using veth pair. Docker0: enables containers to be connected to the host network and other container networks in the sam..

This article is based on the docker seminar on 11, November 2022 from GDSC-Yonsei. I would like to give credit to GDSC Lead for the reciprocal teaching. Problem A python file("hw1.py") runs well on my laptop. But what if it does not on my friend's? ➡ ModuleNotFoundError: No Module named 'xgboost' ➡ The module isn't installed in his laptop Plus, what if there's one who hasn't installed Python? Wh..

Remote Development using SSH · Visual Studio Code Remote - SSH extension allows users to open a remote folder on any remote machine, VM, or container with a running SSH server. Install extension 2. Set up SSH config File Press F1 (inline code shortcut in notion: Ctrl+E) and Click Remote-SSH: Connect to Host… 3. Connect Click the green Open a Remote Window icon on the bottom left corner, then sel..

1. Compiled languages Compiled languages are directly converted into machine code that the processor can execute. The target machine directly translates the program. If a person writes the code, the compiler separates the file and the end result is an executable file. So basically, the owner keeps the source code. Compilation errors prevent the code from compiling. Advantage: tend to RUN faster ..

[ Comprehensive Mapping of Long-Range Interactions Reveals Folding Principles of the Human Genome ] Bona Park Chromosome conformation capture is a way that enables researchers to observe interactions between loci. These loci are in close contact in the 3-dimnesional structure of a chromosome, but they can be apart in the linear sequence. It is important to understand how chromosomes fold that it..

I gave another 40-minute presentation in the Introduction to Bioinformatics class. It was based on the paper "Protein Secondary Structure Prediction Using Cascaded Convolutional and Recurrent Neural Networks". Open in new tab

I gave a 40-minute presentation in the Introduction to Bioinformatics class at the University of Miami. It was reviewing the paper "3D Genome Reconstruction from chromosol contacts" by Annick Lense. It was the first in-person presentation in my university life. 40-minute was a long presentation but I ended up finishing well. New Tab ref: https://www.nature.com/articles/nmeth.3104

Since more and more genome sequence has discovered, it has been popular to predict protein structure from amino acid sequence. So far if protein structure has a template structure, people have predicted an unknown protein structure by comparative modelling. If the applicable protein structure template cannot be found, fold recognition methods have been used. There is also ab initio tertiary meth..

[4강: 분류] - 균일도 기반 규칙 조건 - 정보 균일도 측정 방법 1) 정보 이득 ..앤트로피 개념 정보이득 지수 = 1-엔트로피 지수 2) 지니계수: 불평등 지수 지니계수 낮을수록 균일한 데이터 - 결정트리의 규칙노드 생성 프로세스 If true/ else - 결정트리 장점 쉽고 직관적 - 결정트리 단점 과적합(overfitting) sol) 트리크기를 사전에 제한 - 결정트리 주요 hyperparameter - max_depth, max_features.. - Graphviz이용한 결정트리 모델의 시각화(실제 나무 모양 그림으로) - 각 노드에는 피처의 규칙 조건 gini samples: 현 규칙에 해당하는 데이터 건수 value: 클래스 값 기반의 데이터 건수 ex) [41,4,10] 이면 해당..