A vertical Retrieval based AI conversational agent

A vertical Retrieval based AI conversational agent

Zafarullah Mahmood

Zafarullah Mahmood

New Delhi, Delhi

Design and deployment of a retrieval based AI chatbot ready to be deployed as a product.

Artificial Intelligence

Description

We are designing an AI solution for scheduling meetings. It will be a chatbot that will help its users schedule, reschedule and update meetings with their guests/clients by automating the task. Currently, we are classifying the intent of an utterance using seq2seq and other RNN-based models. The bottleneck we are facing is in named entity recognition and ranking. We are on the hunt for someone who has experience in building chatbots. Our workflow pipeline will be: 1. Intent Classification: Given a mail, classify it based on the sender's intention. E.g., Schedule, Cancel, Reschedule, Update Preference, etc. 2. Entity Extraction: Given a mail, tag each word to a concept E.g., New York -> location, today-> time 3. Response Repository: Templates for various responses 4. Response Ranking: Given potential responses, try to find the best response

Fbh

Zafarullah M. added photos to project A vertical Retrieval based AI conversational agent

Medium 67dc7c27 2e80 47f8 be60 ee85cfb19671

A vertical Retrieval based AI conversational agent

We are designing an AI solution for scheduling meetings. It will be a chatbot that will help its users schedule, reschedule and update meetings with their guests/clients by automating the task.
Currently, we are classifying the intent of an utterance using seq2seq and other RNN-based models. The bottleneck we are facing is in named entity recognition and ranking. We are on the hunt for someone who has experience in building chatbots. Our workflow pipeline will be:
1. Intent Classification: Given a mail, classify it based on the sender's intention. E.g., Schedule, Cancel, Reschedule, Update Preference, etc.
2. Entity Extraction: Given a mail, tag each word to a concept E.g., New York -> location, today-> time
3. Response Repository: Templates for various responses
4. Response Ranking: Given potential responses, try to find the best response

Medium fbh

Zafarullah M. created project A vertical Retrieval based AI conversational agent

Medium 67dc7c27 2e80 47f8 be60 ee85cfb19671

A vertical Retrieval based AI conversational agent

We are designing an AI solution for scheduling meetings. It will be a chatbot that will help its users schedule, reschedule and update meetings with their guests/clients by automating the task. Currently, we are classifying the intent of an utterance using seq2seq and other RNN-based models. The bottleneck we are facing is in named entity recognition and ranking. We are on the hunt for someone who has experience in building chatbots. Our workflow pipeline will be: 1. Intent Classification: Given a mail, classify it based on the sender's intention. E.g., Schedule, Cancel, Reschedule, Update Preference, etc. 2. Entity Extraction: Given a mail, tag each word to a concept E.g., New York -> location, today-> time 3. Response Repository: Templates for various responses 4. Response Ranking: Given potential responses, try to find the best response

Bigger fbh
  • Projects 1
  • Followers 0

Zafarullah Mahmood

Computer Science student at Jamia Millia Islamia, New Delhi

Jamia Nagar, Ghaffar Manzil Colony, Jamia Nagar, Okhla, New Delhi, Delhi 110025, India

No users to show at the moment.