Stanford Seminar - Natural Language Processing for Production-Level Conversational Interfaces

TL;DR

Learn how to build enterprise-level conversational interfaces and improve your machine learning models for better user experiences.

Transcript

thanks Andy good afternoon everyone my name is Karthik raghunathan and this is sarah she ragu and she we are both part of the WebEx intelligence group at Cisco this is the group that works on building machine learning driven intelligent experiences for all of Cisco's WebEx portfolio of collaboration products we are both really excited to be here as... Read More

Key Insights

• 😮 Conversational interfaces are becoming increasingly important as voice searches and connected smart devices continue to rise in popularity.
• ✋ Building a high-quality conversational assistant is a complex task that requires careful data collection, natural language processing techniques, and ongoing model improvement.
• 🖐️ Features like domain classifiers, intent classifiers, entity recognizers, entity resolution, and language parsers play crucial roles in building a conversational assistant.

Questions & Answers

Q: What is the WebEx Intelligence Group at Cisco working on?

The WebEx Intelligence Group is focused on building machine learning-driven intelligent experiences for Cisco's WebEx collaboration products.

Q: What is the goal of the WebEx assistant?

The goal of the WebEx assistant is to improve collaboration within meetings by helping users find the right participants, scheduling meetings, and providing briefing notes and post-meeting summaries.

Q: How can machine learning models be improved over time?

Machine learning models can be improved through ongoing model improvement, using techniques like active learning to select and annotate relevant data for training. This helps improve model accuracy over time.

Q: How does the WebEx assistant handle voice inputs?

The WebEx assistant currently resides in Cisco's video conferencing devices and does not have a mobile app. It uses automatic speech recognition (ASR) to convert audio into text and then processes the text using natural language processing techniques.

Summary & Key Takeaways

• Cisco's WebEx Intelligence Group is working on building machine learning-driven intelligent experiences for WebEx collaboration products.

• The talk focuses on building a complex conversational assistant for the enterprise, covering topics like natural language processing and machine learning models.

• The WebEx assistant aims to make collaboration within meetings more efficient and productive by leveraging AI technologies.