This project is my dissertation project.

Generative AI like ChatGPT, GPT-4, and Google Bard (PALM2) are reshaping various sectors by mimicking human-like responses, blurring the line between AI-generated and human content.

This shift has led to reliance on these AI models for various tasks such as academic writing, fake reviews, misleading news, and social media posts worldwide. To tackle this, multilingual models have emerged to distinguish between human and AI-generated text. However, most prior studies focused primarily on English, with limited testing on other languages like Japanese, German, and Hindi.

Seven models and a perplexity-based method were analyzed, with five models consistently tested across multiple languages. The absence of comprehensive datasets for these languages required new dataset development. Generally, these models performed well when trained and tested on diverse topics but struggled when exposed to single-topic datasets, particularly RoBERTa.

 Misclassifications occurred, with some machine-generated texts being labeled as human-written. BERT showed better overall performance in languages like German and English, while XLM-RoBERTa and DistilBERT-Multilingual excelled with Hindi texts. Perplexity-based methods like GPTZero effectively differentiated between human and machine-generated English texts, suggesting the use of watermarking algorithms by language models. Models specifically pretrained on languages like HindiBERT and BERTJapanese accurately classified human-written and machine-generated text in their respective languages.

For more details visit:

https://github.com/surajjeoor/Human_vs_machine_Analysis

In this project we are fetching the twitter comments related to animal rights. We use tweepy for this particular operation to fetch out the data from the twitter. After fetching this data, we did some data cleaning like normalizing, text cleaning.

The language Models we use over here was Vader, RoBERTa(Robust Bidirectional Encoded Representation Transformer). In this model, we were feeding data record by record, text by text, Models like RoBERTa predicts the sentiment for every text. Those sentiments were recorded.

Some word samples were provided to humans to predict and take out their accuracy. From the report, it became pretty clear that Roberta out performed vader and humans.

This data was provided to power bi to do some analysis, and it turns out that nearly 80-90% of the twitter users are supporting the animal rights.

And also it turns out that the number of vegans on twitter are as much as number of the people in the favour of animal rights.

I have done this project along with my friend so all in all, it was group project.

 I initiated the project of building the image classifier that classifies images of forests from cities. 

This project was all about building image classification model that identifies whether the given image is the forest or city.

As a part of the dataset, I took both forest and street views from the cities, like 

1. Pune, India

2. Accra, Ghana

3. Stirling, United Kingdom

I tried our datasets on lots of pre-trained computer vision models, for instance, VGGNet, AlexNet, and ResNet. Surprisingly, these models gave very low accuracy for our datasets.

Post that, we tried Yolo v5 model on our datasets. The model performed well on my datasets.

Further, I built my own customized convolutional neural network and tried the same dataset on that, which turned out fantastic. The performance parameters of this model were satisfactory. 

For reports and readings, kindly visit the link:

https://github.com/surajjeoor/ITNPAI_Project_Forestcityclassifier

Central Bureau of Investigation(CBI) wanted to expand their databases of the cases with a brand new website.

For this, the team approached Mastek for the solution where Mastek provided the same. This project was divided in two phases.

In Phase 1, the development team designed the portal where each employee of CBI can go and Add, delete, Read and retrieve cases.

The Second phase was all about migration of the data from the oracle database to Microsoft SQL database hosted on the azure and building a proper welcoming page.

I was involved in second phase of this project where I engineered and sustained a high-performing web portal with DotNetNuke, enhancing user experience and functionality. Plus, I expertly troubleshot and resolved backend issues in C# .NET MVC, significantly boosting system reliability.

I spearheaded the seamless migration of historical data from Oracle to Microsoft SQL Server, ensuring data integrity and operational continuity.

In emergencies like flooding, highway incidents, or urgent homelessness, citizens need 24/7 help. Out-of-hours partnership with Capita and Ealing Council, established in 2018, ensures all calls are answered. Capita approached Mastek Limited to develop a solution where customer advisors log complaints via TK Dialog, feeding data to Dynamics 365 CRM on Azure. Each council uses customized SharePoint sites to manage consultant shifts, and cases are assigned accordingly.

In this project, I integrated TKDialog, SharePoint, and Dynamics 365 CRM with custom Dot Net plugins and Microsoft Flow on Azure. I developed Power BI reports, coordinated change requests, and bug reports with project managers, and regularly interacted with clients. I managed a Jira portal with a 100% issue resolution rate, conducted testing with Selenium WebDriver and Cucumber, and received the 'Portfolio Excellence Award' for optimizing productivity and resolving a Power BI daylight saving issue, boosting client profits by 10%. I was also praised by Hammersmith and Fulham council for resolving Azure Active Directory issues.

Our large pool of customer service experts and advanced technology eliminates the need for costly in-house call handling. Our advisors provide personalized, 24/7 support, enhancing service efficiency and resilience, even in crises.

To know more about this project, please click on below link:

https://www.capita.com/expertise/government-services/local-government-services/citizen-experience/emergency-out-of-hours

For more details visit: https://www.academia.edu/32697143/PUBLISHED_OVERVIEW_OF_DRONE_BASED_AGRICULTURE_pdf