talkAItive’s story starts in 2017 when a group of data professionals came together to build a vision where people and institutions are able to understand the big picture and connect the dots to build deeper relationships through emotional contexts.
Our research showed a clear opportunity to learn emotional contexts, which differed from social engagement metrics offered by leading social listening companies.
We quickly ran into the problem of detecting sarcasm and language usage-based contexts which required a different approach from the word dictionary-based analysis.
We found that usage makes all the difference. For example, “killer app” vs “killer dog” communicates contrasting emotional intensity but "killer" in isolation can only be negative.
In 2018, the team achieved our first research success through detection of sarcasm in Twitter chatter which was peer-reviewed by the academic community.
This gave us the early foundation and path to develop novel methods that learns the usage of language by context.
Our focus for research and product development is to identify these contexts related to place, person, product or brand and to measure the sentimental and emotional intensity across time.
"A method and system for dynamically generating a sentiment community based on varied online content sources using a linguistic sentiment engine. A novel approach accounts for a sarcasm sentiment by detecting same and accounting for its influence information of the transient sentiment community. Applicable to identifying and capturing sentiment communities such as for product reviews, service reviews, political and non-political commentary, and other content".
Our patent has given us the focus and the path to become the world’s best emotional understanding technology company.
We have extended talkAItive’s abilities to detect language that uses a variety of phrases to convey the same "emotional context".
This helps our clients quickly summarize the strength of trends, and track narratives across different time periods. Clients can even tell talkAItive to track specific narratives and monitor narratives that sticks vs fizzles out.
This patent was awarded in February 2021 and we have extended it to include novel ways talkAItive detects and monitors emotional intensity across written and spoken word and soon in images.
This journey so far would not have been possible without a team that actually cares about this vision, and are personally motivated to make it a commercial success.
Our learning as a group has been phenomenal, and our discussions cover everything from algorithm design, deep learning networks, NLP, user experience, and application of our insights.
We are constantly trying to add new ways to learn emotional intensity in digital media. If you have related ideas across any of these disciplines, please drop us a line at firstname.lastname@example.org!
I want to help companies and governments better understand the emotional value of their fans and critics, in an open and transparent way. Our understanding of each other can always get better and lead to a great human experience, that is the core belief at talkAItive!
Vibhu is the CEO at talkaitive. He has over 15 years of experience in building technology solutions, leading high-performance delivery teams, and providing consultation at executive level for major Canadian banks. Vibhu has worked with CIBC, Toyota Canada, ATB, and Financial Institutions in the Middle East.
Vibhu believes in hiring right and developing a team-first, supportive culture. He has helped CIBC Commercial Banking develop a culture of innovation, technology adoption, and design thinking. Most recently, Vibhu has been busy building an AI/ML practice focused on NLP and bringing talkAItive to market.
Vibhu holds FRM designation from GARP and has other Financial Services certifications like CSC, Value Investing, and Technical Analysis.
I joined talkAItive to introduce a new form of analytics to Marketing and Business users using AI and ML-based techniques which add value to the existing "Voice Of Customer" business flow. I think it's an exciting time to be part of a team that provides sentiment, emotion, and contextual information to business leads as part of their continuous learning flow. I look to introduce a symbiotic relationship between AI/ML and real business outcomes.
Ravi joined talkAItive in 2017 as a data engineer. His background is in big data computation with expertise in the Apache Spark framework. In 2017, Ravi focused on building the data pipelines for talkAItive using cutting edge data computation techniques. His most valuable contribution has been in migrating talkAitive’s models from Spark to TensorFlow containers. These TensorFlow containers, with the support for GPU based infrastructure, are the backbone of Hiwave’s flagship talkAItive product.
Recently, his focus has been on refining and applying cutting edge techniques to run talkAItive’s deep learning capabilities and analyze millions of data points across hundreds of user themes in real-time. His dream is to be able to talk to talkAItive about any topic - where talkAItive helps him learn something he never knew before.
He is fluent with Scala, Java, Python, Shell script, SQL languages, and building learning on TensorFlow, Spark containers and using backends like Neo4J, MongoDB and PostgresSQL.
Ravi has a Masters specializing in Big Data computation techniques.
I wish to create a product and an experience for businesses to really understand what their fans want in an easy, streamlined way. This would not only provide great feedback for companies to act upon, but would make consumers like myself feel my voice is heard. This passion helps drive me to create the best product that I know this team can. I truely believe talkAItive can make a difference.
Joe joined talkAItive as a co-op student. The ease with which Joe picks up new technology stacks was instrumental in getting our first generation of models off the ground in 2016. Since then he has been working to build models in Spark, building data pipelines using Kafka, enabling API layers using next-generation openWhisk framework, and many more exciting technologies.
Recently, Joe has been building talkAItive’s HTML5/Angular responsive interface. This interface sits on top of openWhisk based API layer, which can dynamically scale to handle traffic of over 100 TPS.
Joe is fluent in Scala, Java, Python, HTML5/Angular/JS, and building learning modules running on TensorFlow and Spark frameworks using hybrid backends like graph, document, key/value or relational.
Joe is a graduate of the University of Waterloo with a degree in Computer Science and a minor in Statistics.
Supriya has been pushing the boundaries of deep learning for the last 2 years at talkAItive. Her core focus has been building intelligence within the realm of natural language processing. She led the effort around building the first high accuracy sarcasm detection in micro-blog data, which was published and nominated for the most innovative idea at the 3rd International Conference On Computer Science Networks and Information Technology in Montreal in 2017.
Recently, her focus has been building talkAItive’s deep learning capabilities which support conversation context generation, media data summarization capabilities, and abstractive inference of how social communities feel about relevant topics. She won’t stop until talkAItive can analyze and advise on topics that impact our clients.
She is fluent in building ML/NN using Python and a related set of extensive libraries.
Supriya has her Masters and is a Ph.D. candidate, specializing in Deep Neural Networks research.