|
||||
|
2016 Jan 4 - 8 project: Cognitive computing
Page last edited by Per Bækgaard (pgba) 28/12-2017
Week 1 project Cognitive computing apps that can interpret images or answer mails are becoming the norm due to recent advances in deep learning and open sourcing of tools such as TensorFlow. Applying a Word2Vec neural network model makes it possible to learn word embeddings i.e. high dimensional feature vectors in order to infer semantic associations and analogies from raw text, or retrieve shared representations enabling classification of novel images based on their nearest word vector neighbors. Sequential representations are likely to evolve into thought vectors capturing patterns of reasoning that will increasingly be incorporated into autonomous software Learning objectives
Monday January 4 8: Short lecture followed by group work on the assignment : conceptualize a cognitive app based on Word2Vec using a lean business model canvas which is to be uploaded to campusnet as a PDF before lunch - if you feel lost play around with the demo app to figure out what semantic associations and analogies can be inferred from the feature vectors and imagine how this could be utilized in a shared representation combining different modalities or a sequence2sequence neural network conversational model 13: Brief presentations and feedback on selected projects followed by group work 16: Upload group work as a PDF containing business model canvas and user story map to peergrade.io - subsequently review the projects you have been assigned. Tuesday January 5 8: Short lecture followed by group work on the assignment: conceptualize a cognitive app based on Word2Vec illustrating its functionality using a landing page which is to be uploaded to campusnet as a PDF before lunch. 13: Brief presentations and feedback on selected projects followed by group work : translating your landing page concept based on user story map task slices into wireframe iterations using POP 16: Upload group work (PDF with business models canvas, user story map, link to public POP url + screenshot of wireframe storyboard) to peergrade.io - subsequently review the projects you have been assigned. Wednesday January 6 8: Short lecture followed by group work on assignment to be uploaded before lunch: add microinteractions (trigger rules feedback loops modes) which describe what happens between the wireframes in your POP storyboard. 13: Brief presentations of selected projects followed by group work: validate your prototype with at least three other groups, utilize the feedback as input for a revised next iteration, update your storybard adding microinteractions (triggers rules feedback loops modes) which describe what happens between the wireframes. 16: Upload group work to peergrade.io and review the assigned projects Thursday January 7 8: Short lecture followed by group work on assignment to be uploaded before lunch: optimize the wireframes in your storyboard by applying gestalt principles based on flat UI design patterns 13: Brief presentations and feedback on selected projects followed by group work: validate your prototype with at least three other groups, utilize the feedback as input for a revised next iteration, update your storybard wireframes by applying gestalt principles based on flat UI design patterns 16: Upload group work to peergrade.io and review the assigned projects Friday January 8 8: Project presentations (3 mins per group) with live peer review assessing 1. to what extent does the prototype address clearly defined user needs for a specific segment ? 2. to what extent does the prototype as a minimum viable product solve these needs ? For the grading, please use this form. The results will be available later here. 17: Upload final project report fulfilling the above learning objectives for a Word2Vec based smartphone app, including lean business model canvas, user story map, screenshot of storyboard and link to POP prototype, using the ACM article format (half a page per group member, only PDFs accepted as hand-in) Video lectures Lean methods related to UX design prototyping, business modeling and user story mapping UX design for qualitative and quantitive prototyping UX design microinteractions: triggers, rules & feedback UX design microinteractions: loops & modes UX design turning shapes into gestalts UX design translating gestalts into design patterns Slides UX modeling of needs and solutions Literature DL4J http://deeplearning4j.org Thought vectors, deep learning and the future of AI DL4J http://deeplearning4j.org Word2Vec Rare Technologies: Word2Vec tutorial and demo app Jon Kolko: "Design thinking comes of age" (Harvard Business Review, 2015) Ash Maura: "Running Lean", Chapter 1 Meta-principles (free sample) Jeff Patton: "User Story Mapping", Chapter 1 The big picture" (free sample) Laura Klein: "UX for lean startups" Chapter 1 Early validation (free sample) Dan Saffer: "Microinteractions" Chapter 1 "Designing microinteractions" (free sample) |
|||