Machine Learning

PT7 Web is an annotated Portuguese language Corpus built from samples collected from Sep 2018 to Mar 2020 from seven Portuguese-speaking countries: Angola, Brazil, Portugal, Cape Verde, Guinea-Bissau, Macao e Mozambique. The records were filtered from Common Crawl — a public domain petabyte-scale dataset of webpages in many languages, mixed together in temporal snapshots of the web, monthly available [1]. The Brazilian pages were labeled as the positive class and the others as the negative class (non-Brazillian Portuguese).

Parallel sentences in English and French, with mathematical expressions tokenized. The French sentences were extracted from course notes on error-correcting codes authored by Dr. Monica Nevins, University of Ottawa.

The dataset has Gaussian Blobs of varying samples, centers and features.  The number of samples ranges from 500 to 50,000. Similarly, the number of centers varies from 2 to 100, while the number of features varies from 2 to 2048. These different sets of Gaussian blobs can be used for testing clustering algorithms for their scalability and effectiveness. There are two kinds of files inside the compressed sets. Files ending with "_X.csv" consist of datapoints, while the files ending with "_y.csv" represent respective class data.

The dataset is composed of digital signals obtained from a capacitive sensor electrodes that are immersed in water or in oil. Each signal, stored in one row, is composed of 10 consecutive intensity values and a label in the last column. The label is +1 for a water-immersed sensor electrode and -1 for an oil-immersed sensor electrode. This dataset should be used to train a classifier to infer the type of material in which an electrode is immersed in (water or oil), given a sample signal composed of 10 consecutive values.

Annotated image dataset of household objects from the RoboFEI@Home team

This data set contains two sets of pictures of household objects, created by the RoboFEI@Home team to develop object detection systems for a domestic robot.

The first data set was created with objects from a local supermarket. Product brands are typical from Brazil. The second data set is composed of objects from the RoboCup@Home 2018 OPL competition.

This dataset gives a cursory glimpse at the overall sentiment trend of the public discourse regarding the COVID-19 pandemic on Twitter. The live scatter plot of this dataset is available as The Overall Trend block at https://live.rlamsal.com.np. The trend graph reveals multiple peaks and drops that need further analysis. The n-grams during those peaks and drops can prove beneficial for better understanding the discourse.

Most text-simplification systems require an indicator of the complexity of the words. The prevalent approaches to word difficulty prediction are based on manual feature engineering. Using deep learning based models are largely left unexplored due to their comparatively poor performance. We have explored the use of one of such in predicting the difficulty of words. We have treated the problem as a binary classification problem. We have trained traditional machine learning models and evaluated their performance on the task.