Service Communautaire d'Information sur la Recherche et le Développement - CORDIS

FP5

PED-CARE Résumé de rapport

Project ID: IST-2000-25417
Financé au titre de: FP5-IST
Pays: Spain

The Intelligent e-Mail Manager (IeMM) tool

Ped-Care project aimed at smoothing the on-line relationship between learners and teachers in the framework of distance learning education through a new communication methodology and supporting tools.

The Intelligent e-Mail Manager (IeMM) aims to help teachers to improve relationships with students offering facilities for a efficient management of the established communication channels (eMail, SMS, Web or WAP) in the distance learning environment.

This Intelligent Agent captures the messages (questions) that students send to the teachers and analyses their content applying semantic analysis techniques (LSI and neural networks completing with steeming algorithms) in order to select the best answer in real time. The message can be automatically answered without human intervention, can be forwarded to the teacher with some suggestions about possible answers or can be redirected to other departments. Several parameters allow users to control the system behaviour (employed algorithms, certainty level, number of suggestions, etc.).

The semantic analysis techniques allow the system to provide an automatic answer for most common questions or to send them to the teacher offering suggestions about recommended answer. Of course the system captures the teacher answer (one of the list of the suggested ones or a new one) in order to allow the training of the semantic analysis algorithms. For this purpose, the IeMM is connected to knowledge base containing the most frequent asked questions (FAQ) and their associated answers. In this way, when the teacher rejects the suggestions’ list and enters a complete new answer, it is automatically stored in the FAQ allowing the system to provide the right answer next time the same or similar question is received.

One of the advantages that the intelligent e-mail provides for the distance learning environment is that there is no need of being complaint with e-learning platforms or Learning Management Systems (LMS). Moreover, we are conscious that there are a lot of e-mails systems able to filter, collect and sort e-mails but our definition of intelligent agent together with the application of artificial intelligence techniques (semantic analysis for natural language processing, support of any language, segmentation tools based on pedagogical indicators for providing personalised responses) is unique in the market.

Even if the main IeMM potential applications lie on the distance education environment, it is important to highlight its possible applications to other sector from those that it was initially conceived. Within this areas we can mention:
- CRM providers taking advantage of IeMM functionality for their helpdesk services and for market segmentation purposes.

- Knowledge Management System providers.

- e-Business solutions.

- Intranet and Internet corporative Web portals.

- Call Center and Data Centers interested of offering CRM solutions through Internet or Mobile devices.

This tool is also of interest in general, to any organisation interested in offering helpdesk services through Internet (Intranet) to their customers (B2C, B2b) or employees (B2C). It has been developed with the last generation technologies (intelligent agents, AI techniques) offering access by any media at any time (sms, e-mail, wap and web).

The main intelligent e-Mail Manager characteristics are:
- Reception of users' questions by any media (Web, Wap, eMail or SMS)

- Application of Semantic Analysis Techniques (LSI, Neural Networks, Pattern Matching) for analysing received questions in order to select the best set of answer from the Knowledge Base. It also incorporates steaming algorithms for removing "noise" from the received messages

- Configurable system behaviour when questions are received allowing:
--Automatic answering to users selecting the best choice from the selected ones.
-- Redirection of the question to the assigned agent including recommendations of possible answers. Regardiog education, the agent could be the official teacher of the course or any other department of the education centre.
-- Incorporation of the user profile (if available) to the question redirected to the agents providing extended information about his behaviour in the organisation. Regarding education, the learner profile will be attached to the question sent to the teacher including information about his activities in the learning process.
-- Personalised answers to the same question depending on the segment to which the user is assigned. Regarding education, the answer sent to learners could be different for those belonging to advanced segments (for example, those that have had a better progress in the course) than for those that are in the low qualified segments.

- Group Communication facilities, allowing agents to send messages (eMail, SMS or Fax) to the whole group, to specific segments or to selected users. It also possible to establish Personalised Marketing Campaigns (different offers to each population segment) having also Analysis Tools for evaluating the campaign success. Regarding education, the teacher can launch different education offers (for example, seminars) to each student’s segment having the possibility of analysing how many students have shown interest in the activity.

- FAQ list including common questions and answers. The answers support the incorporation of multimedia contents as well as connections to contents searchers. In this way, video in MPEG4 format can be complete the answer sent to the user and also "More Info..." buttons can be added that offer the users other contents from external repositories (see PCM and it facilities of connection with OAI repositories or Ariande's LOM repositories).

- Dynamic FAQ construction. The knowledge base can be directly updated when the agent rejects the system recommendations offering new answers to user. In this way, if the teacher decides that the suggestions offered by the system to one learner question are not adequate, he can write a different answer that is automatically registered in the knowledge base and in the associated FAQ list.

- Multilingual Capabilities. The applied IA algorithms for the natural language processing are based on statistical methods that are independent of the language of the received questions.

- Automatic system training incorporating the agent actions to received questions.

- Reports, statistics and graphics regarding use of the different communication channels.

The requirements for incorporating IeMM within an existing solution are:
- Connection to the corporate Mail Server. IeMM incorporates a Mail Server that receives all the messages. Then, they are sent to the corporate Mail Server to be redirected to the assigned agent.

- Knowledge Base feed with the actual organisation FAQ.

- Initial training process for ensuring an acceptable response from the beginning. Of course, the IA algorithms that IeMM incorporates will improve its performance with the time.

- For offering personalised answer the installation of the IPSM (Intelligent Pedagogical Segmentation Manager) is required.

- If SMS support is required, the Ped-Care SMS Centre is also required.

Regarding the costs for implementing an IeMM solution, it is based on Open Source software so only one server is required with Linux, a Virtual Java Machine and a James Server. Therefore is not necessary to purchase additional software licences. In case that the support to SMS is required and additional server is required to install the SMS Centre (again based on Open Source solutions).

Potential users of IeMM are: Private and Public Universities; Education Centres; e-Learning providers; Human Resources Departments of large companies; Companies that offer e-Business solutions on Internet specially those applying CRM strategies; Call Centers and Data Centers; CRM solution providers; Knowledge Management providers; Public Administration Organisms.

Reported by

Altran SDB
C/ Ramirez de Arellano, 15 6? planta
28043 Madrid
Spain