Recently Updated Pages

Optimizer filter notes

The DAML implementation uses this filter when constructing the optimizer Adam(lr = self.meta_lr, ...

Swahili Syntax (Anthony Vitale, 1981)

Grammatical Sketch Very brief grammatical sketch, strong focus on syntax which is nice to see si...

Dependency Parsing

Info

https://github.com/allenai/allennlp/blob/master/allennlp/tests/training/multi_task_trainer_test.py

DAML walkthrough

The main point of interest is train_maml in model.py. prev_min_loss gets set to a high value on ...

DAML code

import torch import random import numpy as np from config import global_config as cfg from re...

Research questions for thesis

How do Bantu languages, and Swahili in particular, integrate into the Universal Dependencies fra...

F-score for positive class analysis

Bag of Words alone Feature Count Mutual Information Chi-squared 100 0.61468 0.61682 ...

Macro f-score analysis

General sentiment features (MPQA) Feature Count Mutual Information Chi-squared 100 0.59...

Precision and Recall for top performers (macro f-score)

MI features Data Source N-features Negative Precision Negative Recall Positive Precision ...

Methods for building twitter-specific sentiment lexicon

Using mutual information This page has quite a bit of information regarding how the mutual infor...

Comparison of two-side MI

As mentioned in Methods for building twitter-specific sentiment lexicon there are two general way...

Results dump

Cross validation character n-grams tfidf F1-score Task A 0.6440953412784399 preci...

Todo: Coding

Implement sentiment value creator with fit() and transform(). Integrate corenlpSentimentAnalyze...

Log

Fixed issue where the server would not work with the sentiment analysis annotator enabled by upgr...

Design specs

Original Page Layout If generate is selected, the lexical entries and syntactic rules are used...

Gameplan (in order of priority)

Utilities Intake of rules Transformation of non-binary rules into binary rules Generation of ...

Handling Optionality

The issue is that the optional pieces are occurring in rules of any length. E.g. we could have NP...

Hai's notes

upper limit for num of generations: S -> NP VP NP:{John, Mary} VP:{eat, sleep} when I as...

Genie tidbits

css ane other frontend assets should be modfiied in the public folder not the assets/css folder.