L8 and it's interesting use of data

Page 1 of 1 (5 items)
This post has 4 Replies | 1 Follower

Posts 1162
Mike Tourangeau | Forum Activity | Posted: Wed, Jan 23 2019 9:26 AM

I have long thought that Logos has a ton of data (cultural concepts, Sense, various data sets) that could, when used together in an algorithm generate some interesting information.

"Important Passages" is an example of this. Is this a form of artificial intelligence? What would you call this?

My thoughts are: 1. I hope Faithlife continues to pursue this type of data use, I think there is more that could be done to mine our vast libraries. Esp if it was combined with the community. (i.e. - 78% of users spend the most time/highlight/note in this resource when studying this passage. Prioritize these results)

2. Is "Important Passages" more intensive on a processor? My old mac is getting ready to "give up the ghost" but I think it struggles with these type of searches.  I am getting ready to build a new Logos machine for the office. The common wisdom is a SSD is the biggest upgrade you can do for Logos but I am thinking CPU use will become more important in the years to come. Esp if Logos continues to put together things like "Important Passages"

What do you think?

Posts 1485
Forum MVP
Levi Durfey | Forum Activity | Replied: Wed, Jan 23 2019 9:56 AM

Mike Tourangeau:

"Important Passages" is an example of this. Is this a form of artificial intelligence? What would you call this?

Sean Boisen:
Both Important Passages and Important Words/Lemma in Passage are data gathering processes that we periodically run over all the available commentaries (not manual tagging), So, these dataset should continue to be updated as we add more commentaries to Logos.

Mike Tourangeau:
I think there is more that could be done to mine our vast libraries. Esp if it was combined with the community. (i.e. - 78% of users spend the most time/highlight/note in this resource when studying this passage. Prioritize these results)

Important Passages doesn't mine your commentaries. It's something that FL periodically does on their end to all the commentaries available in Logos. Then the result is shipped to users. Therefore, it's not going to be CPU intensive.  

Posts 13065
Forum MVP
Mark Barnes | Forum Activity | Replied: Wed, Jan 23 2019 10:13 AM

Mike Tourangeau:
"Important Passages" is an example of this. Is this a form of artificial intelligence? What would you call this?

It's not artificial intelligence. It's just an human-created algorithm (albeit a complex one).

Machine learning is the next step. All the big firms are spending billions on that at the moment, and it will gradually filter down over the next several years.

Mike Tourangeau:
Is "Important Passages" more intensive on a processor?

Not particularly. With most of these new datasets, all the heavy lifting is done on Faithlife's servers (in this case, when the database that powers this was built). Disk speed remains the single most important factor for us. Probably on Faithlife's servers then CPU speed is much more important.

Posts 25956
Forum MVP
MJ. Smith | Forum Activity | Replied: Wed, Jan 23 2019 2:59 PM

Mark Barnes:
Machine learning is the next step. All the big firms are spending billions on that at the moment, and it will gradually filter down over the next several years

From some comments made by FL employees, I thought they already used training data and, by extension, some AI techniques.

Rick Brannan (Faithlife):
I'll look into this. The classification is automated, but some of the underlying training data for the automation may have changed (or a few other possible areas).

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 1162
Mike Tourangeau | Forum Activity | Replied: Thu, Jan 24 2019 6:17 AM

Very interesting. Thanks for the clarification and info. 

Page 1 of 1 (5 items) | RSS