1 of 6

Rezrv Bot

Simplifying IoT based Automation

Vinamra Palaiya, Chinmayi Appaji, Lokesh Johri

Kausik Lakaraju, Biplav Srivastava

A demonstration at CASY2 on 15^th Oct 2021 https://casy.aiisc.ai

2 of 6

Problem Statement and Background

Customer service is costly and can also be improved with more speed and accuracy technology. There is a general industry trend of introducing decision-support in the networking domain.

Building a conversational agent for decision support around router management makes networking services more accessible to end users and it also accelerates the speed with which issues faced by the users can be resolved.�
We aim to build a state-of-the-art conversational agent which could assist the user by providing information about router usage, in modifying different settings based on user preferences and provide solutions for resolving router issues from a user perspective.�
The solution is situated in a connected environment (like home) with multiple users.

Key Aim: Manage Networked Environments for Users with a Robust Conversational Agent

The novelties of our conversational agent in contrast to existing work are:

Dialogue-level
Use of natural interaction (e.g., voice commands) for user sub-groups (e.g. Parental Control).
Use of multi-speaker authentication using voice control.
Support language variants.
Expedited user onboarding.
System-level
Using small data. E.g., auto-generation of intents, and train with more data when available.
Perform with real-time considerations.
Optimize networking resources when desired.
Have a robust pipeline with configurable/ multiple technology options.

Business Value

Control of usage in networked environment - e.g., home - using natural interaction (voice, gesture).

Bootstrap without data; when available, use dialog/ interaction data to learn and customize response.�
Interoperates with multiple chatbot platforms.

Invention Diagram

References

A. V, A. K H, A. Kamal K and B. Kumaraswamy, "Multi-level Speaker Authentication: An Overview and Implementation," 2020 IEEE 17th India Council International Conference (INDICON), 2020, pp. 1-6, doi: 10.1109/INDICON49873.2020.9342423.

Lin, Z., Liu, Z., Winata, G.I., Cahyawijaya, S., Madotto, A., Bang, Y., Ishii, E., & Fung, P. (2020). XPersona: Evaluating Multilingual Personalized Chatbot. ArXiv, abs/2003.07568.

https://techcrunch.com/2016/10/19/oyoty-is-a-chatbot-designed-to-teach-kids-to-be-safe-online/