MétaCan
Menu
Back to cohort
Record W3112748649 · doi:10.1002/alz.046767

Detecting language impairment using ELIEC

2020· article· en· W3112748649 on OpenAlexaff
Mahboobeh Parsapoor

Bibliographic record

VenueAlzheimer s & Dementia · 2020
Typearticle
Languageen
FieldPsychology
TopicMental Health via Writing
Canadian institutionsUniversity of Toronto
Fundersnot available
KeywordsComputer scienceNatural language processingArtificial intelligenceReadabilityClassifier (UML)StatisticSpeech recognitionMathematicsStatistics

Abstract

fetched live from OpenAlex

Abstract Background Machine learning (ML) can detect language impairment. However, it is challenging, if we do not have sufficient language data. To deal with this issue, we suggest using ELiEC or "Emotional Learning‐inspired Ensemble Classifier". It is a new ML algorithms that can learn from a few examples. This paper briefly describes ELiEC (structurally and functionally) and presents preliminary results obtained from employing it on the textual datasets extracted from speech produced by 14 (i.e., five patients and nine healthy control) subjects. We collected the speech samples via a web portal named Talk2Me. Method We have developed a language assessment tool combining natural language processing (NLP) techniques and the ELiEC (as an ML tool). NLP techniques extract the total number of word tokens and the total number of unique word types, the total number of sentences, and the total number of syllables. Using the above parameters, we calculate lexical features such as the lexical diversity score, Brunet’s Index (BI), Honore’s Statistic (HS) , Flesch‐Kincaid, and the Flesch Reading‐Ease (FRES) Test readability scores. ELiEC maps linguistic features to patients and healthy controls. In more detail, the ELiEC, which is an ensemble classifier, is developed based on LeDoux’s emotional theory. The theory describes the neural structures underlying the processing of threatening stimuli (see Figure 1) We develop ELiEC by combining classifiers (see Figure 2) according to the interconnections between regions of the brain that are responsible for processing threatening stimuli. Result Table 1 compares results obtained from using various ML tools such as Support Vector Machines (SVMs), k nearest neighbor, and ELiEC to distinguish patients and healthy subjects. As observed, ELiEC’s accuracy is 0.87 (+/‐ 0.33). Thus, compared to other ML tools, it can associate language impairment to people with dementia with the highest accuracy and when there is not a sufficient amount of language data from patients. Conclusion Our results verified that by utilizing ELiEC, we could develop an accurate MLbased language assessment tool when there is a limitation in recruiting patients. We will improve the ELiEC framework and use it to distinguish people with different types of dementia.

Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.

How this classification was reachedexpand

Full frame distilled prediction

Teacher imitation

Not calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.

metaresearch head score (Codex)0.000
metaresearch head score (Gemma)0.000
Version: codex-gemma-dda1882f352aValidation status: machine_predicted_unvalidated
Candidate categoriesInsufficient payload (model declined to judge)
Consensus categoriesInsufficient payload (model declined to judge)
DomainCandidate signal: none · Consensus signal: none
Study designCandidate signal: Other design · Consensus signal: none
GenreCandidate signal: Empirical · Consensus signal: Empirical
Teacher disagreement score0.712
Threshold uncertainty score1.000

Codex and Gemma teacher scores by category

CategoryCodexGemma
Metaresearch0.0000.000
Meta-epidemiology (narrow)0.0000.000
Meta-epidemiology (broad)0.0000.000
Bibliometrics0.0000.000
Science and technology studies0.0000.000
Scholarly communication0.0000.000
Open science0.0000.000
Research integrity0.0000.000
Insufficient payload (model declined to judge)0.0040.001

Machine scores (provisional)

The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.

Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.

Opus teacher head0.074
GPT teacher head0.373
Teacher spread0.298 · how far apart the two teachers sit on this one work
Validation statusscore_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from it

Classification

machine, unvalidated

Machine predicted; both teacher heads agree on what is shown here.

Study designOther design
Domainnot available
GenreEmpirical

How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".

Quick stats

Citations5
Published2020
Admission routes1
Has abstractyes

Explore more

Same venueAlzheimer s & DementiaSame topicMental Health via WritingFrench-language works237,207