MachineLearnIT
This post compares ways we can import packages to Python. Our old friend, the pip and the requirements.txt duo works fine until it doesn’t, so new packaging tools address the dependency management of package import as well. This post compares pipenv, pip-tools and poetry.
Continue reading “Comparisons of pipenv, pip-tools and poetry”
XLNet: Generalized Autoregressive Pretraining for Language Understanding by Yang et al. was published in June 2019. The article claims that it overcomes shortcomings of BERT and achieves SOTA results in many NLP tasks.
In this article I explain XLNet and show the code of a binary classification example on the IMDB dataset. I compare the two model as I did the same classification with BERT (see here). For the complete code, see my github (here).
There are some innovations in IT that everyone follows and hear about and then there are some others, quieter, humbler, those sort of innovations that hide behind a wall and when you understand their power and what you can do with them, you are truly astonished.
Git is the most widely used version control system in the world. I have to use it during my work, and as it is not difficult to use (if you don’t have conflicts), I didn’t think much about it until recently. However, once I started to read more about it, I quickly realised its elegant way of dealing with versions and integrity control worth more than an article! We have probably all taken this amazing tool granted time to time, underestimating all the trouble it saves us from. So let’s realise the power and elegance of git, and let’s finally get into a committed relationship with it!
Continue reading “Let’s have a committed relationship … with git”
One of the major advances in deep learning in 2018 has been the development of effective NLP transfer learning methods, such as ULMFiT, ELMo and BERT. The Transformer Bidirectional Encoder Representations aka BERT has shown strong empirical performance therefore BERT will certainly continue to be a core method in NLP for years to come.
Continue reading “BERT: Bidirectional Transformers for Language Understanding”
During one of my missions, I could compare two database management systems and their performance on text lookups: Elasticsearch and PostgreSQL. Here is the what I found.
The aim of this article is to have an introduction to Naive baysian classification using scikit-learn. The naive Bayesian classification is a simple Bayesian type of probabilistic classification based on Bayes’ theorem with strong (so-called naive) independence of hypotheses. In this article, we will use it to build a basic text prediction system. We will predict Equity codes in a search form fashion (i.e prediction starts when user starts typing).
Continue reading “Equity codes prediction using Naive Bayesian Classifier with scikit-learn”
This article explains SentencePiece, a language-independent subword tokenizer and detokenizer introduced by Kudo et al., 2018 and implemented in Python and C++. SentencePiece implements two subword segmentation algorithms, the Byte-Pair Encoding (BPE, Sennrich et al., 2016) and the Unigram language model (Kudo et al., 2018).
During my job I have to analyse and correct mathematical models that aim to evaluate risk in C++ and automise processes in Python. As I come from a finance background, I thought it could be a good exercise to look at some elementary data structure and code them from scratch in these two programming languages. Here comes the version C++.
