Meeting Minutes from AI Lab session on Saturday 15th June at Bengaluru
#CellStratAILab #disrupt4.0 #WeCreateAISuperstars
CellStrat AI Lab in engaged in incredible AI innovations and product development activity. The sophistication of our AI Lab members’ presentations continues to rise week after week.
Last Saturday AI Lab started with my session on Decision Trees. Decision Trees are powerful ML models that can be used both for classification or regression. We train Decision Trees with the CART algorithm wherein we try to minimize the weighted average of impurity of two child nodes at each parent node split. Decision Trees and their variation Random Forests might overfit the training data, for which we resort to regularization of the trees by stopping the tree from splitting beyond a certain depth.
Then came an intense and rich healthcare presentation on RNA-RNA (ribonucleic acid) analysis by Jidhu Mohan. This model discovers interaction between different RNA sequences of human body using Deep Neural Network algos such as CNN or LSTMs.
Ribonucleic acid (RNA) is a polymeric molecule essential in various biological roles in coding, decoding, regulation and expression of genes. RNA binding protein play a crucial role in gene regulation. Experimental approaches for detecting the protein binding sites on RNAs are time consuming and costly. In this case study, deep learning methods were used to predict RNA-RNA interaction from the sequencing based training data. An RNA-based RISE dataset was used for modeling and validation.
Then came a superb session on Abstractive Text Summarization by Indrajit Singh, where the model produces a compressed re-wording of the original text. This allows one to save time that would be spent in reading long form text. This finds applications in media, legal, healthcare, business reporting and variety of other content heavy industries.
Indrajit presented a specialized technique for abstractive text summarization which is based on semantic understanding of the original text. This is called Rich Semantic Graph (RSG). Here a semantic graph is produced (using domain ontology data) from original text (RSG creation), condensed to a smaller graph (RSG reduction) and then a language summary is produced from the latter graph. RSG is a new ontology-based technique and finds uses in machine translation, text summarization and information retrieval.
Wish to learn AI and ML at India’s No 1 AI Lab ? If yes, join us for the Saturday AI Lab Meetup on 22nd June 2019 at BLR :-
Register : https://www.meetup.com/Disrupt-4-0/events/261132379/
Topic : Image Segmentation w/ UNET, DNN innovations for Vision
Date : Saturday 22nd June 2019, 10:30AM – 5:00PM
Loc. : WeWork, Embassy Tech Village, ORR, BLR
See you on 22nd June for the AI Lab meetup !
Questions ? Call me at +91-9742800566 !
Co-Founder & Chief Data Scientist, CellStrat