Natural Language Processing and Its Application in the Utility Industry

Virtual Breakout Session

Thursday, June 11, 2020
11:00am EDT
Loading Sessions

Register now!

This is a consume-at-your-own-pace event, it will take place 100% online.  All of the content has been pre-recorded and will be released at its scheduled time. If you can’t listen or watch a session when it’s released, remember that you can access it at ANY TIME in the next 30 days.

  • Listen to podcasts
  • View webinars
  • Watch videos
  • Access articles, white papers, case studies and research reports
  • Chat with speakers and attendees via discussion forums
  • Visit sponsor profile pages

As the need to automate many processes grows significantly, tools such as natural language processing (NLP) and text mining are more utilized in the utility industry. NLP has great potential to derive valuable insights from many forms of data available to the industry, especially textual data. These insights may help the industry reduce costs and improve safety. This session will include discussion around the fundamentals of NLP and use cases in the power industry.

Feel free to ask questions in the discussion forum below - speakers will be responding as quickly as they can. View this webinar by the end of the day on June 11 and get entered to win a $100 Amazon Gift Card.

Speakers:
Sponsored by:

Discussion

4 Comment threads
5 Thread replies
0 Followers
 
Most reacted comment
Hottest comment thread
newest oldest most voted
Martin Richardson
Member
Martin Richardson

Do you provide dictionaries for Gas/Electric industries? Or are there some available in industry? How do these handle Canadian differences?

Sanam Mirzazad
Member
Speaker
Sanam Mirzazad

Our focus is on the Electric Industry, and as far as I know, there is no dictionary available in the electric industry for text mining and NLP applications. The process should be that different, but the algorithm needs to be trained on relevant data, which shows the difference.

Martin Richardson
Member
Martin Richardson

How many digs do not call 811? What portion of line hits does the “un-ticketed” group make up?
FortisBC is pursuing this currently. Please call if you would like to discuss.

Xiwang Li
Member
Speaker
Xiwang Li

Thank you for your message. I will share your information with our product manger.

Duminda Randeniya
Member
Duminda Randeniya

@Sanam: Thank you for a very good presentation. One of the problems we face is multiple acronyms. Can you explain how you tackle these multiple acronyms in developing the dictionary? Are developing a mapping from each acronym to respective (human intervention) word or use POS-tags to do this (100% machine intervention)?

Sanam Mirzazad
Member
Speaker
Sanam Mirzazad

Thanks, Duminda. We are doing a combination of both; use algorithms (PoS-tagging, n-grams, etc.) to identify abbreviations and their different variations. Then we use human intervention to map them to respective words. This part can also be automated to some extent, using glossaries that we have in the utility.

Steve Lach
Member
Steve Lach

How much should a human be involved in a process which is based on NLP in the power industry?

Xiwang Li
Member
Speaker
Xiwang Li

Hi Steve, Great question. NLP, like any other machine learning or AI field, we need human be involved in different stages. Like we need human to label/annotate the data, we need human to test/evaluate the model/tool. So now for power industry, we still need human effort in these stages.

But hopefully, as the model gets stronger and stronger, we many need less and less of human efforts.

Sanam Mirzazad
Member
Speaker
Sanam Mirzazad

It depends on the size of the available data. More data means less human intervention. However, for an industry-specific dictionary, a human needs to be present in the loop and cannot be eliminated.

Related sessions

No related sessions currently available