Cognitive Factory Module

{Note : This is a preliminary documentation in progress, while dev taskforce is working}

History

The Cognitive Factory modle is a final rework of the Cognitive Factory prototype designed for Moodle 1.9, subsequently redraw of the Brainstorm module.

Basic Concepts

Most human skills are often reductible to a set of cognitive primitives that you must be trained on for achieving tasks in the real world. A cognitive primitive is an operation that we can execute on a set of inputs. We talk about cognitive operations when input are ideas or predicates in respect to computational operations where inputs are calculable values.

The Cognitive Factory module proposes to exploit those operators to build kind of exercices that cannot be built with standard quizes, as proposing a complex mental operation to be performed, maybe combining several operators in the same exercice.

The Cognitive Factory allows each instance to organize an exercice combining some operators and designing a workflow with those operators.

Organisation

The Cognitive Factory module is modular and provides a set of cognitive operators that can be extended by developers.

Exercise Workflow

The module is organized to run a workflow defined by the teacher, running through 5 steps (plus a final grading step). Each step of the workflow can be assigned to either a student or the teacher which multiplies the flexibility of the Cognitive Factory. When changing such step affectation the global behaviour of the exercice can change and address new pedagogic situations in the audience.

Here are described the five workable steps of the cognitive exercise process :

Phase 1: Ideas collection

Users provide textual predicates (ideas) to the factory. This might be done f.e. during some among of time, before going on next steps. Ideas can be fed in a "single user scope" process (each user bringing ideas in his own space whithout collective visibility) or in a group brainstorm process, sharing ideas with the group.

Phase 2 : Operations selection and setup

Once a pool of ideas has been collected, and enough stuff is available to "think" with, some applicable operators have to be choosen in order to build the cognitive exercice. Again, this can (usually) be done by the teacher, but there is sense the learners be asked to make this choice in some activity configuration. At the moment, this choice will be global to the activity instance.

Phase 3 : Operation execution

The users are asked to use the operators to process the initial set of ideas and propose an output. This outcome is always individual, but might be influenced by less or more information visible over what other people is doing aside.

Phase 4 : Outcome display and confrontation

In this phase is assumed sufficiant data has been input for results exploitation and matching or non macthing to make sense. The Display phase attemps to give users a panorama over complete responses which scope depends on overal or local privacy settings. The purpose of display layouts is to help some conclusion to emerge from the initial rough set of predicates.

Phase 5 : Feedback and synthesis

This is the last activity step that asks to participants to write a motivated synthesis over experimentation and outcoming conslusions. This step is reduced to a simple online text editor where to write report.

Extra phase : Evaluation

When the module is used in an evaluated process, a grade can be elaborated by graders to participate to the gradebook compilation. The evaluation only can be used by teachers (graders). In the evaluation process can a feedback be posted by teachers to users, that will display in the synthesis panel.

Direct Workflow and Reverse Workflow

We will talk about direct workflow not necessarily for a student application to the task, but for the most probable and trivial assignation of the task.

For example, in a brainstorm activity, the trivial assignation of the first step is asking students to collect ideas. We will call this the direct flow. Operators are usually designed to let students perform the operation that has been cofigured by teachers. We'll call this the direct flow again. When teacher performs the operator, this will be the reverse flow. There is a master direct flow (what the cognitive factory was initally designed for), and several reverse flows possible.

To summarize :

• Master direct flow
  • Student collects
  • Teacher sets up operators (defines work)
  • Student perform operations
  • Student analyses answer
  • Student reports about observations
  • Teacher feedacks to student or collectively
  • Teacher grades work done

• Reverse flow (example of)
  • Teacher collects
  • Student sets up operators (defines work)
  • Student performs operations
  • Teacher analyses answers
  • Teacher feedacks to student
  • Teacher grades work done

Pedagogic situations regarding results privacy

Working with cognitive puzzles is known being strongly affected by interrationality factors. This means the result of a cognitive exercise will not be the same wether the appliant thinks alone, or is more or less influenced by other people in the neighbourhood.

The Cognitive Factory operators integrate this variation of the exercise objectives, by providing three major configurations regarding privacy of operations.

When executing operators as "blind", appliant will have no information at all about the behaviour of peers when processing the operator requirement. They will proceed the operator in a string mono-rationality, without external influence. Some statistics can be accessed on Display phase.

When executing operators as "semi-blind", appliants can have some information about the peers attitude, such as global trends, that can affect the answer they give to the puzzle.

When executing in "full view", the appliant has full information upon the peer results.

In addition, a global "privacy" switch alters the whole activity behaviour adding one privacy level :

• Isolated: Appliants only use and see their own inputs (or teacher's inputs if the collect phase is assigned to teacher). No access at all to peer results (student inputs), or limited access to global trends (teacher inputs).
• Blind: Appliants play with all inputs. Appliants WILL NOT SEE any peer information before having processed the operators once.
• Informed: Appliants play with all inputs. Appliants CAN SEE trends and statistics before they have processed by their own.

Two sessions of the same cognitive exercice could easily be setup to demonstrate this effect to students.

The privacy beahaviour is summarized by the following tables:

For input collection:

COLLECT Isolated I Blind I>C Informed C

Legend: I : Individual answers C : Collective (Group or Course) > : Second operation possible only if first is achieved

For operator processing :

PROCESSING Blind Semi-Blind Full view Isolated P(I) P(I) P(C)>T Blind P(C) P(C)>T P(C)+T Informed P(C)>T P(C)+T P(C)+T+A

Legend: P : Processing GUI P(I) : Processing only own answers P(C) : Processing all answers T : Trends and audience stats A : Anonimized individual answers table > : Second operation possible only if first is achieved

For results exploration:

DISPLAY Blind Semi-Blind Full view Isolated D(I) D(I) D(C)+T Blind D(C)+T D(C)+A D(C)+F Informed D(C)+A D(C)+F D(C)+F

Legend: D : Display (own results) D(I) : Display only results on own answers D(C) : Display on all answers T : Trends and audience stats A : Anonimized individual answers table F : Full individual answers table

Operators

• Description template
• Categorizing
• Ordering
• Quantifying
• Reduction by elimination
• Reduction by synthesis
• Hierarchization
• 2D Mapping
• Dependencies Analyis and Searching Cycles
• Graphing

Credits & Team

Cognitive Factory is presented by the MyLearningFactory Team.

Author : Valery Fremaux (valery.fremaux@gmail.com) Contributors : Wafa Adham (Adham.ps Ltd).