| The FreeHAL application is the most advanced self-learning artificial intelligence available as free (open source) software. The FreeHAL project was founded in 2006 by Tobias Schulz, and is focused on the further development of this artificial intelligence. The application uses semantic networks and employs pattern recognition, stemming, part-of-speech databases and Hidden Markov Models in order to best mimic human behavior in conversations. In contrast to most free, as well as commercial chatbots, FreeHAL is able to add to its own knowledge. The program expands its knowledge base through typed communication with the user. Both the German and English languages are supported, but at the moment an extensive database - in fact, a semantic network - exists only for German. In the current development version available for download, the database encompasses 110,779 words and 203,532 facts. FreeHAL supports an online version which you can try without installation. The FreeHAL team is looking for enthusiasts who are willing to participate actively in FreeHAL. More detailed instructions can be found on the participation page   !The subproject FreeHAL@home is used to distribute computer databases to as many computers as possible, in order to convert texts that are found on the Internet into facts that can be used by the main project. Specifically, this process involves semantic networks, which we plan to include in the official download in the medium term. Basically, all data obtained via BOINC is put after a certain time online. Before this can be done, the data is tested by sampling, we review the ability of FreeHAL to use the facts to reach significant conclusions, and create reports about the success rates of the Tagger and Parser components, enabling us to find and resolve bugs. At the moment, articles in the German and English Wikipedia are converted into semantic nets, and later we plan to use other data sources such as the free projects ConceptNet and OpenCyc. Similarly, free databases such as OpenThesaurus will be converted into the FreeHAL format. These processes are amenable to distributed computation, making BOINC a natural choice. For the Part-of-Speech Tagger, Hidden Markov Models and deterministic statistical data are used. |
| Name | # | Users |
|---|---|---|
| Lobby | 0 | |
| FreeHAL | 0 | |
| BOINC | 0 |