The "INDIAai" (National AI Portal of India) portal provides weekly comprehensive articles highlighting the research contributions made by universities and colleges in India.

We aim to offer thorough reporting on the AI research contributions made by a single institution every week. This series allows researchers and students to provide concise explanations of their work.

RV College of Engineering (RVCE) is a prestigious self-financing engineering college established in 1963. It is affiliated with Visvesvaraya Technological University (VTU) Belagavi and operated by Rashtreeya Sikshana Samithi Trust (RSST).

Let us explore the top AI research contributions from RV College of Engineering, Bengaluru, India.

Temple Inscriptions Recognition and Transliteration in Devanagari Script 

The authors B. Sathish Babu, Sannidhi Shetty, Anushka Agarwal, Sai Lahari Sreerama, Vaishnavi K Bhustali, and Sanka Sanjana states that, "Ancient inscriptions, palm scripts, manuscripts, etc., have vital information about India's rich culture. Recognition and understanding of these inscriptions have been challenging for epigraphers and professionals."

They further said, "The proposed research aims to advance optical character recognition methods for archival Vatteluttu script inscriptions, which date back to the 4th or 5th century AD. This paper discusses a deep learning model to transliterate the ancient Tamil inscriptions (Vatteluttu Script), which can be extended further to other languages. The proposed work benefits epigraphists, archaeological researchers, and the general public interested in this topic."

The authors further explained, "The work uses Neural networks to extract information to recognize ancient Tamil scripts. The outcomes of the prehistoric character identification task using the CNN model demonstrate that accurate categorization is feasible when employing deep neural networks. By using a larger dataset and an enhanced neural network method for recognition, this system can overcome the problems plaguing the earlier designs. The developed deep learning model has achieved an accuracy of 84.12%."

The prototype of this project was awarded the Smart India Hackathon 2022 Software Edition award and published in IEEE Xplore.

Automated government form-filling for aged and monolingual people using an interactive tool  

The authors are Adarsh R. Hegde, R. S. Sujala Reddy, P. Kruthika, B. C. Pragathi, Sree Rama sai lahari, N. Deepamala and G. Shobha.

While explaining the project, the authors state, "The Government of India offers various schemes for various classes of citizens. Applications of the schemes are to be filled in English, and monolingual individuals find it difficult to access and fill the forms." 

The authors further express, "The project tries to solve the challenges faced by monolingual individuals in India, particularly the elderly, people with impairments, and those from marginalized communities. The proposed work is to create an interactive "Dhvani" Voicebot specifically designed for the Kannada language. It helps users identify suitable government schemes and fills forms in English."

Furthermore, they added that the project was envisioned to help the aged and needy by filling out forms in post offices and banks. 

The prototype of this project was completed and received second prize in Lab2Marker 2023. This research is published as an open-access article in the Taylor and Frances journal. 

Automated dynamic schema generation using knowledge graph 

The authors of this research work are Priyank Kumar Singh, Sami Ur Rehman, Darshan J., Shobha G., and Deepamala N.

The authors say that "Database is utilized in almost all the fields. Database developers find it difficult to manage, store and query huge amounts of available data. Establishing schema relationships and enabling developers to filter, prioritize, and recommend appropriate schema is essential."

They added, "Database developers get tailored schemas and services by browsing through a vast volume of dynamically produced schemas with a recommendation system." 

The authors said, "Many machine learning algorithms require more time and data to address problems. In this project, Knowledge graphs are used to solve the problem. A huge knowledge graph is built, and the schemas are extracted and recommended based on a query on Natural Language."

This project uses the latest Knowledge Graphs to solve the automated schema generation problem. The project led to a publication in the IAES International Journal of Artificial Intelligence (IJ-AI).

Natural language to structured query language using ElasticSearch for descriptive columns 

The authors Sourabh S Badhya, Akshar Prasad, Shetty Rohan, and Y S Yashwanth state that "Today, numbers, text, photos, etc. are captured in massive amounts. The data captured is very complex to generate information." 

They added, "Data is stored in relational databases like MySQL and PostgreSQL or document-oriented databases like MongoDB and Cassandra. Extracting data from these databases requires special knowledge of writing queries designed for that particular database." 

The authors said, "The project extracts information from input natural language query and converts it into a query the database can understand. This work uses Elastic search to extract data from descriptive columns."  

This project was implemented to help database users to query the database using natural language. The project is published in IEEE Xplore.