The description of this case study in Dow and the MOLTO website makes apparent that the math use case demonstrator is not so much a translation editor as natural language front end to computer algebra.
Leader: jordi.saludes
Timeline: July, 2010 - May, 2012
The ultimate goal of this package is to have a multilingual dialog system able to help the math student in solving word problems.
The UPC team, being a main actor in the past development of GF mathematical grammars and having ample experience in mathematics teaching, will be in charge of the tasks in this work package with help from UGot and UHEL on technical aspects of GF and translator’s tools, along with Ontotext on ontology representation and handling. We will start by compiling examples of word problems. In parallel, we will take the mathematical multilingual GF library which was developed in the framework of the WebALT project and organize the existing code into modules, remove redundancies and format them in a way acceptable for enhancement by way of the grammar developer’s and translator’s tools of work packages 2 and 3 (D6.1). The next step will be writing a GF grammar for commanding a generic computer algebra system (CAS) by natural language imperative sentences and integrating it into a component (D6.2) to transform the commands issued to the CAS (Maybe as a browser plugin). For the final deliverable (D6.3), we will use the outcome of work package 4 to add small ontologies describing the word problem: We will end with a multilingual system able to engage the student into a dialog about the progress being made in solving the problem. It will also help in performing the necessary computations.
The impression is confirmed by an email From Jordi Saludes:
"The simplest implementation will be a terminal-based question/answer
system like ELIZA, but focused on solving word problems. It will start by
giving the statement of the problem, then it will do computations for the
student/user, list unknowns, list relations between unknowns, state the
progress of the resolution and, maybe, give hints.
We are thinking about the kind of word problems which require solving a system
of (typically two) linear equations. In Spain these are addressed to first or
second year high school students."
On the way to the demonstrator, the plan is to devise small ontologies describing math word problems and verbalise them using the MOLTO platform and WebAlt project math GF grammars. These phases of the work can be evaluated on the lines indicated under WP2-3. Since the corpus is small, manual quality evaluation using TAUS criteria is appropriate. We need to buy TAUS criteria if we are not getting them from the patent partner.
ID |
|
Task leader |
Status |
New comments |
6.0 |
Hold |
|
||
6.1 |
Planned |
|
||
6.2 |
Planned |
|
||
6.3 |
Ongoing |
|
||
6.4 |
Planned |
|
||
6.5 |
Planned |
|
||
6.6 |
Planned |
|
||
6.7 |
Planned |
|
||
6.8 |
Planned |
|
ID |
|
Due date |
Dissemination level |
Nature |
Publication |
D6.1 |
1 June, 2011 |
Public |
Prototype |
|