Tous les articles

ALISE – A brief overview – Part 1

Architecture and Functions At the center are the 4 main modules of the semantic mind of ALISE: Aglaia, Hephaistos, Peripatos and Leto. When we start from the top Aglaia is the interface manager. Wi…

Architecture and Functions

At the center are the 4 main modules of the semantic mind of ALISE: Aglaia, Hephaistos, Peripatos and Leto.

When we start from the top Aglaia is the interface manager. With all her grace (after all she is one of the three Graces) she communicates between the Al!se components and the outside word; translates between JSON, XML/RPC (until v2) and the program.

Calls are secure transmitted via JSON Web Token (JWT – RFC 7519). You must verify a signature before storing and using it. The Web Token is returned from the server after successfully logging in, stored in the Clients Session Store and deleted after terminating the session or a defined time (so it is necessary to request a new token if the session is terminated).

Hephaistos, the god of smithing, forges the information gathered from the other modules and controls the communication between them.

Further down we have Peripatos, the place of the great philosophical school of Aristoteles and unsurprisingly where the brain of the program is located. Here the AI uses statistical methods of probability and analyzes words and sentence structures. The text is further enriched with semantic metadata.

The information is then saved and organized in the database Leto. The Database manages long term knowledge and delivers case based data. The ontologies, according to which the information is characterized, are customized depending on the specific needs of the program.

On top of the main modules is Selene, the interface. Selene is browser based and therefore accessible anywhere. It is connected to the harvester and crawler Athena, which scans and analyzes given documents and collects data. The main function of Selene is the dialogue/chat where questions are answered based on the learned information (scanned in Athena, connected in Hephaistos, analyzed in Peripatos and stored in Leto). The semantic AI will be able to interpret the previously input data and more importantly understand its meaning. According to the information available, it can now make decisions and answer questions.

Aglaia – sample JSON Calls

Calls are secure transmitted via JSON Web Token (JWT – RFC 7519). You must verify a signature before storing and using it. The Web Token is returned from the server after successfully logging in, stored in the Clients Session Store and deleted after terminating the session or a defined time (so it is necessary to request a new token if the session is terminated).

Aglaia Calls

Method Request Body Response
Authentication In authentication, when the user successfully logs in using their credentials, a JSON Web Token will be returned and must be saved locally (typically in local storage, but cookies can be also used), instead of the traditional approach of creating a session in the server and returning a cookie. Requested parameters are um which is a valid eMail address of the user, password which is mandatory to receive a token but if not set rqn has to be set with the parameter un to generate a new account. If password is not set and rqn+pwd the password recovery is initiated starting with a mail to the user. POST http(s)://[aglaiahost]://api/v1/login {“user”: “user@mail”,”pwd”: “password”} {“fn”:”firstName”, “ll”:”lastName”, “sk”:”base64securekey”)Or, if unsucessful {}

Hephaistos Calls

Getter Method Request Body Response
Get available semantic DataBases POST http(s)://[aglaiahost]://api/v2/get { “sdb”: “”, „token“:token} Returns all available ontologies or {}
Get available ontologies POST http(s)://[aglaiahost]://api/v2/get { “ont”: “”, „sdb“: sdburi, „token“:token} Returns all available ontologies or {}
Get specific ontology POST http(s)://[aglaiahost]://api/v2/get { “ont”: “semcare”, „sdb“: sdburi, „token“:token} Returns ontology or {}
Get specific class with attributes+datatype, sliks+sliksattribs and subclasses POST http(s)://[aglaiahost]://api/v2/get { “ont”: “semcare”, „class“:“[classuri]“, „sdb“: sdburi, „token“:token} Returns class of ontology with attributes+datatype, slinks+slikattrib and subclasses-uri
GetSOb: get semantic object POST http(s)://[aglaiahost]://api/v2/get { „soburi“:“[soburi]“, „sdb“: sdburi, „token“:token} Returns requested object with all set attributes and slinks.
GetSBOb: get binary content of semantic object POST http(s)://[aglaiahost]://api/v2/get { „soburi“ : “[soburi]“, bin : true, „sdb“: sdburi, „token“ : token} Returns binary content of requested sOb.
Get all sOb fitered by class POST http(s)://[aglaiahost]://api/v2/get {Filter[“class, attrib, slink-value_0…N”], „sdb“: sdburi, „token“ : token} Returns all sOb if values=nul, otherwise intersection of values depending on filteroperator.
Setter Method Request Body Response
SetSOb: set Semantic Object POST http(s)://[aglaiahost]://api/v2/set { “SObURI”:”[URI],”SObLabel”:”[LABEL], “token”:token} Returns {“SOBURI”:”[SObURI]”, “State”:”[ErrorMessage]”} – ErrorLevel 0=ok
SetSObA: set Attributes of a specific SOb POST http(s)://[aglaiahost]://api/v2/set {}
setSLink: link two SOb with a SLink POST http(s)://[aglaiahost]://api/v2/set {}
Delete Method Request Body Response
DelSOb: delete Semantic Object POST http(s)://[aglaiahost]://api/v2/del { “SObURI”:”[URI],”SObLabel”:”[LABEL], “token”:token} Returns {“SOBURI”:”[SObURI]”, “State”:”[ErrorMessage]”} – ErrorLevel 0=ok
DelSObA: delete Attributes of a specific SOb POST http(s)://[aglaiahost]://api/v2/set {}
delSLink: delete SLink of two SOb POST http(s)://[aglaiahost]://api/v2/set {}

Peripatos Calls

Method Request Body Response
POST http://desktop-v7g1qdl:11000/api/v2/question {“text”:”The doctor gives the judge a bottle of wine.”,”user”:”mp”,”model”:”de”,”sid”:”1″} {“text”: “I think… The act has to do with ‘giving'(urn:uuid:5d64d7fc-7bb4-48ea-a5f6-2f4b6fc1058b)”, the person involved ‘doctor'(urn:uuid:a18871c9-1e37-4e32-802c-0b032eadeb13)(perpetrator) and ‘wine’ (new word)(victim/object) to do”}

Articles liés

The Power of AI

Easy ways to work with data like humans do Creating AI supported solutions needs efficient data management, which is object oriented and elastic to give smart algorithms the capability to remember …

1 min de lecture

From Data to Meaning: The Power of Semantic Telemetry in AI

Semantic Telemetry is changing how we understand user behavior — by showing the meaning behind every interaction. Instead of just tracking system data, this approach uses large language models (LLM…

1 min de lecture

Préférences relatives aux témoins

Choisissez les technologies optionnelles autorisées. Vous pouvez modifier ces paramètres à tout moment.