2019/10/26
First Phase - 80%
- Front-end program can extract sentence from speech with high accuracy
- Currently no parameters or specific actions can be determined
Python speech to text → Part-Of-Speech tagger
TODO
- Extract of device name (NOUN type)
- Start of second phase
2019/11/2
First Phase - 100%
- Actions can be determined
- Back-end analysis can extract
DEVICE NOUN
from text
TODO
- Continue onto second phase
2019/11/9
Second Phase – 40%
- Added Syntatic Dependency Parsing allowing creation of parse trees
- Separation of
DEVICE NOUN
,PARAMETER NOUN
and their correspondingVERB
- Elimination of stopwords
Python speech to text → Part-Of-Speech tagger → Syntatic Dependency Parser → Rule Matching
TODO
- Rule-based matching to account for inaccuracy in parse trees
- Word embeddings(BERT/Flair) for synonyms
2019/11/16
Second Phase – 60%
- Started revising output format to a 4 element tuple
- Research about how to handle if statement
TODO
- Finish the 4 element tuple
- Re-dig into SpaCy for different command structure
2019/11/23
Second Phase – 85%
- Synonym partially done
- New output format basically done with some minor errors to deal with
TODO
- Add robustness of strcture analysis
- Wrap off handling synonym & acronym
2019/11/30
Second Phase – 100%
- Synonym & acronym handling done
- Robustness enhanced
TODO
- Start doing research for the stretch goal
- Extra effort on system robustness
- Handle if statement
2019/12/7
Third Phase Head Start
- Research about third phase
- If statement partially done
TODO
- Final presentation preparation
- Finish if statement
- Weblite update
- Final video