國立虎尾科技大學 |

Enhancing Mental Health Diagnosis With Representation Learning : = A Transformer Based Approach.

紀錄類型:	書目-語言資料,手稿 : Monograph/item
正題名/作者:	Enhancing Mental Health Diagnosis With Representation Learning :/
其他題名:	A Transformer Based Approach.
作者:	Kieback, Adrian I. C.
面頁冊數:	1 online resource (83 pages)
附註:	Source: Masters Abstracts International, Volume: 86-02.
Contained By:	Masters Abstracts International86-02.
標題:	Mental health. -
電子資源:	click for full text (PQDT)
ISBN:	9798383627600

Enhancing Mental Health Diagnosis With Representation Learning : = A Transformer Based Approach.
Kieback, Adrian I. C.

Enhancing Mental Health Diagnosis With Representation Learning :A Transformer Based Approach. - 1 online resource (83 pages)

Source: Masters Abstracts International, Volume: 86-02.

Thesis (M.S.)--San Diego State University, 2024.

Includes bibliographical references

Depression is a significant global health concern, affecting an estimated 280 million people worldwide, which not only impacts individual well-being but also imposes a substantial economic burden on societies. This research leverages the Distress Analysis Interview Corpus - Wizard of Oz (DAIC-WOZ) to advance depression detection using two novel neural network architectures: GAAMAudioNet and CustomAttentionTransformer. Both models incorporate the Gaussian Attention Adaptive Module (GAAM), a custom attention mechanism designed to improve the interpretability of deep learning models. The two models enable precise identification of the frequency bands that are most indicative of depressive states, facilitating advancements in the field of explainable AI (XAI). This capability allows for a clearer understanding of model decisions, aligning with the goals of creating more transparent AI systems. Notably, both models are optimized for computational efficiency, featuring significantly lower parameter counts-280k for GAAMAudioNet and 1.1 million for CustomAttentionTransformer-compared to other high-performing models, which often exceed 33 million parameters. The reduced complexity does not compromise their effectiveness; both models demonstrate diagnostic accuracy in detecting depression from audio data. This approach not only enhances the accuracy of depression detection but also ensures that the underlying computational models remain interpretable and manageable in terms of resource utilization, paving the way for broader application and further research in AI-driven healthcare diagnostics.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798383627600Subjects--Topical Terms:

564038
Mental health.
Subjects--Index Terms:

DepressionIndex Terms--Genre/Form:

554714
Electronic books.

Enhancing Mental Health Diagnosis With Representation Learning : = A Transformer Based Approach.
LDR:02943ntm a22003857 4500 001 1150535
005 20241028051516.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798383627600
035 $a (MiAaPQ)AAI31491379
035 $a AAI31491379
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Kieback, Adrian I. C. $3 1477028
245 1 0 $a Enhancing Mental Health Diagnosis With Representation Learning : $b A Transformer Based Approach.
264 0 $c 2024
300 $a 1 online resource (83 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Masters Abstracts International, Volume: 86-02.
500 $a Advisor: Elkins, Aaron.
502 $a Thesis (M.S.)--San Diego State University, 2024.
504 $a Includes bibliographical references
520 $a Depression is a significant global health concern, affecting an estimated 280 million people worldwide, which not only impacts individual well-being but also imposes a substantial economic burden on societies. This research leverages the Distress Analysis Interview Corpus - Wizard of Oz (DAIC-WOZ) to advance depression detection using two novel neural network architectures: GAAMAudioNet and CustomAttentionTransformer. Both models incorporate the Gaussian Attention Adaptive Module (GAAM), a custom attention mechanism designed to improve the interpretability of deep learning models. The two models enable precise identification of the frequency bands that are most indicative of depressive states, facilitating advancements in the field of explainable AI (XAI). This capability allows for a clearer understanding of model decisions, aligning with the goals of creating more transparent AI systems. Notably, both models are optimized for computational efficiency, featuring significantly lower parameter counts-280k for GAAMAudioNet and 1.1 million for CustomAttentionTransformer-compared to other high-performing models, which often exceed 33 million parameters. The reduced complexity does not compromise their effectiveness; both models demonstrate diagnostic accuracy in detecting depression from audio data. This approach not only enhances the accuracy of depression detection but also ensures that the underlying computational models remain interpretable and manageable in terms of resource utilization, paving the way for broader application and further research in AI-driven healthcare diagnostics.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Mental health. $3 564038
653 $a Depression
653 $a Healthcare diagnostics
653 $a Machine learning
653 $a Deep learning
655 7 $a Electronic books. $2 local $3 554714
690 $a 0800
690 $a 0347
690 $a 0769
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a San Diego State University. $b Microbiology. $3 1413449
773 0 $t Masters Abstracts International $g 86-02.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31491379 $z click for full text (PQDT)