Pinned Repositories
air-hockey-with-unity-ml-agents-reinforcement-learning
In the project, I started off by doing a Humming Bird tutorial from Unity-Learn to get an idea about ML-Agents, reinforcement learning, and training a model. Simultaneously, I got the chance to explore the Unity Editor and tried out building the entire scene. While training the bird’s model, I understood a lot about hyperparameters which I then tweaked and tuned to try out new models, the trial and testing sometimes gave horrible results making me realize how much a change in a parameter could affect the performance of a model. Once I was through the tutorial and had a basic understanding of how ML-Agents work, I decided to build other games and empower them with ML-Agents. I chose Air Hockey where wrote my own observations and change the config file to train the model in such a way that it performs best. Initially, I was lost, but then gradually figured out how to train the player similar to the hummingbird. It didn't work out great at the beginning, then I added our own observations, rewards, made some code changes, and provided a scoring UI to finally create an amazing ML-Agent that was smart enough to never let us win!!
amazon-database-system-plsql
The project implements the Amazon's e-commerce database system which fulfils all the functional requirements for an e-commerce website. The database is created by designing the Extended Entity Relation (EER) model of Amazon as an E-commerce website. EER is then converted to relational tables using the set of well defined rules and after applying the normalisation techniques. Project utilizes PLSQL and it's stored procedures, triggeres and cursors to create tables and maintain the consistency within the relational database.
beer-wine-distribution-ecommerce-website
Beer and Wine Distribution E-commerce website
context-based-video-seek-google-cloud-intelligent-video
The project gives the ability to seek videos to a particular point/portion based on the key context provided by the user. Based on a user query (object, genre, category, etc.) the project filters out the videos that have the key context in them and also highlights the portion in the video that where it appeared. The project utilizes Google's Cloud Video Intelligence, Cloud Speech-to-Text API, Google Cloud Storage via Google Cloud Platform, and using Node.js for the integration of everything.
k-means-clustering-tweets-from-scratch
The program clusters simmilar tweets using KNN algorithm from scratch without libraries utilizing the jaccard distances.
lyft-data-challange
Given a large data set of lyft riders, riders are grouped into several categories based on the number of featues using optimal machine learning strategy.
neural-net-breast-cancer-prediction-from-scratch
A neural network (NN) having two hidden layers is implemented, besides the input and output layers. The code gives choise to the user to use sigmoid, tanh orrelu as the activation function. Prediction accuracy is computed at the end.
object-detection-Faster-CNN
This project is a work of fiction written from the perspective of a 2020 researcher traveling back in time to mid 2013 to share some 2020 xNNbased application ideas; references to credit the actual inventors of the various ideas is provided at the end
RNN-based-new-lyrics-generation-from-scratch
In the project, the aim is to generate new song lyrics based on the artist’s previously released song’s context and style. We have chosen a Kaggle dataset of over 57,000 songs, having over 650 artists. The dataset contains artist name, song name, a link of the song for reference & lyrics of that song. We tend to create an RNN character-level language model on the mentioned dataset. Using model evaluation techniques, the model is checked for its accuracy and is parameter optimized. The trained model will predict the next character based on the context of the previous sequence and will generate new lyrics based on an artist’s style.
sentiment-analysis-of-streaming-tweets-and-visualizations-using-its-kafka-kibana
Implemented the following framework using Apache Spark Streaming, Kafka, Elastic, and Kibana. The framework performs SENTIMENT analysis of hash tags in twitter data in real-time. For example, we want to do the sentiment analysis for all the tweets for #trump, #coronavirus.
Ashwanikumarkashyap's Repositories
Ashwanikumarkashyap/k-means-clustering-tweets-from-scratch
The program clusters simmilar tweets using KNN algorithm from scratch without libraries utilizing the jaccard distances.
Ashwanikumarkashyap/amazon-database-system-plsql
The project implements the Amazon's e-commerce database system which fulfils all the functional requirements for an e-commerce website. The database is created by designing the Extended Entity Relation (EER) model of Amazon as an E-commerce website. EER is then converted to relational tables using the set of well defined rules and after applying the normalisation techniques. Project utilizes PLSQL and it's stored procedures, triggeres and cursors to create tables and maintain the consistency within the relational database.
Ashwanikumarkashyap/sentiment-analysis-of-streaming-tweets-and-visualizations-using-its-kafka-kibana
Implemented the following framework using Apache Spark Streaming, Kafka, Elastic, and Kibana. The framework performs SENTIMENT analysis of hash tags in twitter data in real-time. For example, we want to do the sentiment analysis for all the tweets for #trump, #coronavirus.
Ashwanikumarkashyap/air-hockey-with-unity-ml-agents-reinforcement-learning
In the project, I started off by doing a Humming Bird tutorial from Unity-Learn to get an idea about ML-Agents, reinforcement learning, and training a model. Simultaneously, I got the chance to explore the Unity Editor and tried out building the entire scene. While training the bird’s model, I understood a lot about hyperparameters which I then tweaked and tuned to try out new models, the trial and testing sometimes gave horrible results making me realize how much a change in a parameter could affect the performance of a model. Once I was through the tutorial and had a basic understanding of how ML-Agents work, I decided to build other games and empower them with ML-Agents. I chose Air Hockey where wrote my own observations and change the config file to train the model in such a way that it performs best. Initially, I was lost, but then gradually figured out how to train the player similar to the hummingbird. It didn't work out great at the beginning, then I added our own observations, rewards, made some code changes, and provided a scoring UI to finally create an amazing ML-Agent that was smart enough to never let us win!!
Ashwanikumarkashyap/context-based-video-seek-google-cloud-intelligent-video
The project gives the ability to seek videos to a particular point/portion based on the key context provided by the user. Based on a user query (object, genre, category, etc.) the project filters out the videos that have the key context in them and also highlights the portion in the video that where it appeared. The project utilizes Google's Cloud Video Intelligence, Cloud Speech-to-Text API, Google Cloud Storage via Google Cloud Platform, and using Node.js for the integration of everything.
Ashwanikumarkashyap/neural-net-breast-cancer-prediction-from-scratch
A neural network (NN) having two hidden layers is implemented, besides the input and output layers. The code gives choise to the user to use sigmoid, tanh orrelu as the activation function. Prediction accuracy is computed at the end.
Ashwanikumarkashyap/object-detection-Faster-CNN
This project is a work of fiction written from the perspective of a 2020 researcher traveling back in time to mid 2013 to share some 2020 xNNbased application ideas; references to credit the actual inventors of the various ideas is provided at the end
Ashwanikumarkashyap/beer-wine-distribution-ecommerce-website
Beer and Wine Distribution E-commerce website
Ashwanikumarkashyap/lyft-data-challange
Given a large data set of lyft riders, riders are grouped into several categories based on the number of featues using optimal machine learning strategy.
Ashwanikumarkashyap/mobile-net-v1-modified-CNN
This project is a work of fiction written from the perspective of a 2020 researcher traveling back in time to late 2012 to share some 2020 network design, training and implementation of MobileNet V1, references to credit the actual inventors of the various ideas is provided at the end
Ashwanikumarkashyap/Rialto
A marketplace for university students that provides a localized buy-and-sell environment that you can trust.
Ashwanikumarkashyap/RNN-based-new-lyrics-generation-from-scratch
In the project, the aim is to generate new song lyrics based on the artist’s previously released song’s context and style. We have chosen a Kaggle dataset of over 57,000 songs, having over 650 artists. The dataset contains artist name, song name, a link of the song for reference & lyrics of that song. We tend to create an RNN character-level language model on the mentioned dataset. Using model evaluation techniques, the model is checked for its accuracy and is parameter optimized. The trained model will predict the next character based on the context of the previous sequence and will generate new lyrics based on an artist’s style.
Ashwanikumarkashyap/task-assigner-client
A simple not so formal JIRA - A task assigner for the members of the family.
Ashwanikumarkashyap/task-assigner-server
A simple task assigner among the registered members of the family.
Ashwanikumarkashyap/utd-course-viewer
A portal for the students of The University of Texas at Dallas, Computer Science Department where one can check the list of available courses. The database is updated realtime.
Ashwanikumarkashyap/utd-course-viewer-admin
UTD Course Viewer (Admin control)
Ashwanikumarkashyap/decisiontree-model-online-ad-prediction
Decision Tree buliding (without Library) and application of appropriate pruning strategy to an advertisement dataset to increase the accuracy of predicted final classes (purchased/not purchased) by 5%.
Ashwanikumarkashyap/less-disracting-youtube
Using the Youtube Data API v3, the projects fetches only the top five results based on the user's search query and display it to the user. The User can play from the results and search for more as he/she likes. No recommended videos are displayed to maintain the project's simplicity.
Ashwanikumarkashyap/virtual-auction-house