A Single Place for all Pandas Window Functions — Window functions are very powerful in the SQL world. However, there isn’t a well written and consolidated place of Pandas equivalents. Basics of writing SQL-like code in pandas covered in excellent detail on the Pandas site. However, the Pandas guide lacks good comparisons of analytical applications of SQL and their…

Semantic search is a hard problem worth solving in NLP. — We commonly spend a lot of time looking for a specific piece of information in a large document. And we commonly find if using CTRL + F. The proverbial Google-fu, the art of effectively searching for information on google is a valuable skill in a 21st-century workplace. …

Have you ever had to summarize a lengthy document into key points? Or providing an executive summary to a document? …

How to obtaining High-Quality Labeled datasets with Croud-Sourced Workers As a data scientist, we spend an ungodly amount of time handling data — cleaning, normalizing, labeling. These days, thankfully, many solutions off-load the labeling to third parties, freeing up data scientists’ valuable time, and lessening the burden of manual text, pictures or video labeling. However, as a data scientist…

How to get up and running with XLNet and Pytorch in 5 mins — What is XLNet XLNet is a modern NLP language model that is based on Transformers (BERT, RoBERTa, TinyBERT, etc.) Results of XLNet on various Natural Language Understanding tasks are approaching that of human performance. XLNet can generate text at a level of a high-schooler, it can answer simple questions. It can comprehend that…

Overview of data labelling for AI, new paradigms, and size of the growing data labelling market. — Data Labeling Landscape With the emergence of AI, many firms coming to realize the real bottleneck to bootstrapping Machine Learning is a lack of labelled datasets. In response to the need, many companies emerged that offer labelling services, labelling platforms or some other labelling solutions such as providing specific domain-experts to label data…

How to quickly extract text and data from Municipal Bond CAFR Reports What is Spreading A large portion of finance is dedicated to writing and reading financial statements. In the US for a financial statement to be considered official, it must be in a format of PDF. …

How to accurately process PDF files with OCR-as-a-service from AWS Asynchronous API responses Recently a new paradigm of async API responses has become prominent. It works by returning Job-ID rather than an API response. Then, to check the status, the user would need to submit a second call to the API with the…

Overview of two major multilingual sentence embedding models Why Multilingual Models Multilingual Models are a type of Machine Learning model that can understand different languages. One example would be to classify whether a piece of text is a toxic comment. Using a regular Machine learning model we would be able to detect only…