News for hackers. Stuff that matters.

srogatch

Re:A game recommender engine like Akinator on Saturday January 26, 2019 @01:56PM (#19007760)
The database isn't hard-coded. The game information like name and URL is stored in SQL DB. The statistics is stored in engine's knowledge base. Currently only admins can enter games to the system. You can suggest a game by sending me an email: [email protected] with "ProbQA" in the subject.

Ask HN: Can you suggest a game not yet in our interactive recommendation engine?

We are now filling up the game base for http://probqa.com/ - an interactive game recommendation engine based on probabilistic question-asking system (ProbQA). In each step it asks the user about a game feature and offers multiple choices. The engine also lists top 10 recommended games in each step. Recommendations change as the user answers more questions. It's tolerant to errors in the user...

( Read More... | no comments | on HN )
Re:A game recommender engine like Akinator on Saturday January 19, 2019 @06:21PM (#18949861)
An application of the probabilistic question-asking system is now available as a website: http://probqa.com/ . Here users can find the next game to play without knowing its name or keywords. The users only need to answer questions, and the program lists the top recommendations for each user.

The engine that powers the website is supposed to work similarly to a popular game Akinator, where the user thinks of a character, and the program asks questions to guess about the user's secret character.

In ProbQA there is no secret: the user simply doesn't know exactly what he/she wants. So the program asks questions to come up with something suitable for the user.

Please, let me know your thoughts about the website and the engine.

Re:A way to infer software architecture from source code [pdf] on Thursday February 02, 2017 @10:02AM (#13550937)
Thanks, but I'm looking for investments in the project.
Eventually artificial intelligence software must be able to analyze its own source code in order to improve itself. This document describes a first step: a program that analyzes the source code of other programs (and potentially, itself) to infer the software architecture as a nested decomposition of the analyzed software system into subsystems.