Payment reliability is the most important metric for the Lightning Network as a payment system.

We teamed up with @HelloStillmark to deliver research that leverages machine learning to bring enterprise payment reliability to the Lightning Network.

Our New Research: "Channel Balance Interpolation in the Lightning Network via Machine Learning" decimates the tired and risky approach of probing to find a reliable payment route.

Combining crowdsourced data and machine learning means reliable payments on LN with less spam.

We're applying this methodology to a new pathfinding-as-a-service feature that is showing extremely promising early results.

We're looking for exchanges, rewards programs, or play-to-earn with high outbound payment volume to put our pathfinding service to a real-world test.

To learn more about our research or to join us as part of our payment operations solution, be sure to visit: https://rpo.dev/pathfinding

This research would not be possible without our collaborators:

@HelloStillmark

@vsingh_5

@emaros96

And peer reviewers:

@renepickhardt

Dr. Christian Kummerle

@alexbosworth

Reply to this note

Please Login to reply.

Discussion

Love this nostr:npub1clk6vc9xhjp8q5cws262wuf2eh4zuvwupft03hy4ttqqnm7e0jrq3upup9

OFAC Compliance?

This research is about payment reliability, which has a broad range of applications.

You could apply this to risk management/compliance but that's a very small application compared to payment reliability as a while.

Initial results are showing payment reliability increasing from 27% with native LND to 90% using pathfinding derived from predictions.

We're looking for folks to battle test it for real world payment reliability especially for high volume business applications. Link: rpo.dev/pathfinding

nostr:nevent1qqsryr7084shrsgucc0ps2hs0urle42yz4k2jdpk3g4althf2dhlhaqpz3mhxue69uhhyetvv9ujuerpd46hxtnfdupzq2hsrcxkh5delwu784p32ltyty8mylw0hj479puy5kpju9a7lwrmqvzqqqqqqyavuqvz

Please don’t copypasta your tweets here. Go the extra step to tag or untag.

These folks don't have nostr profiles to tag, but they're still deserving of credit!

Our pathfinding performance research begins today!

Using our machine learning predictions, participants will be able to generate a comparison between their node's default pathfinding and our enhanced pathfinding.

These results will be used to further improve payment reliability on the Lightning Network and deliver more research about decentralized payment channel networks.

If you'd like to participate in our benchmark research and access our advancements in pathfinding and payment reliability, be sure to sign up here:

https://rpo.dev/pathfinding

nostr:nevent1qqsryr7084shrsgucc0ps2hs0urle42yz4k2jdpk3g4althf2dhlhaqpz9mhxue69uhkummnw3ezuamfdejj7q3q9tcpurtt6xulhw0r6sc404j9jraj0h8me2lzs7z2tqewz7l0hpasxpqqqqqqzcc2vm8