Implementation of a RAG-LLM Contextual Agent for iVR Learning
Video overview
System aims
- To allow learners to initiate a chat session and pose questions using their
voice. - To allow for both textual and verbal responses from the LLM.
- To allow user control for LLM speech to be paused and replayed.
- To allow for integration with a customised Retrieval Augmented Generation
(RAG) model for persistent context and knowledge. - To incorporate context within the LLM prompt that will give the LLM adhoc
context depending on the game state or specific object the learner is
asking about.
System design diagrams
Component overview
The system makes use of an external LLM (Eden AI) to process queries. This was chosen due to its flexibility and ease of use, facilitating easy testing of different models. A local system was not considered, since the application is for live teaching and integrating with standalone headsets and unsupervised use is crucial. It also integrates with Wit.ai to provide dictation and text to speech (TTS) functions.
Context and response flows
The RAG system contains base chatbot instructions and context, with activity-specific instructions, and object-specific context prepended to the student’s query.
Article references
5189002
{5189002:HXU3RYB5},{5189002:MGRNGC8L},{5189002:6YJQENBX},{5189002:NKT7XGAZ},{5189002:VN3R262U},{5189002:RASV3NUR},{5189002:ZCZGX2AY},{5189002:WJINJTK6},{5189002:6DRDGIID},{5189002:VAW4G3BU},{5189002:4IC89ZTV},{5189002:KPPH5FNM}
1
apa
50
default
1294
https://www.nicolafern.com/wp-content/plugins/zotpress/
%7B%22status%22%3A%22success%22%2C%22updateneeded%22%3Afalse%2C%22instance%22%3Afalse%2C%22meta%22%3A%7B%22request_last%22%3A0%2C%22request_next%22%3A0%2C%22used_cache%22%3Atrue%7D%2C%22data%22%3A%5B%7B%22key%22%3A%22HXU3RYB5%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Ding%20and%20Chen%22%2C%22parsedDate%22%3A%222025%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BDing%2C%20S.%2C%20%26amp%3B%20Chen%2C%20Y.%20%282025%29.%20%26lt%3Bi%26gt%3BRAG-VR%3A%20Leveraging%20retrieval-augmented%20generation%20for%203D%20question%20answering%20in%20VR%20environments%26lt%3B%5C%2Fi%26gt%3B.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-ItemURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fwww.semanticscholar.org%5C%2Fpaper%5C%2Faf54e8314d03df54d1e1857096b053692e325cbc%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fwww.semanticscholar.org%5C%2Fpaper%5C%2Faf54e8314d03df54d1e1857096b053692e325cbc%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22RAG-VR%3A%20Leveraging%20retrieval-augmented%20generation%20for%203D%20question%20answering%20in%20VR%20environments%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Shiyi%22%2C%22lastName%22%3A%22Ding%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Ying%22%2C%22lastName%22%3A%22Chen%22%7D%5D%2C%22abstractNote%22%3A%22Recent%20advances%20in%20large%20language%20models%20%28LLMs%29%20provide%20new%20opportunities%20for%20context%20understanding%20in%20virtual%20reality%20%28VR%29.%20However%2C%20VR%20contexts%20are%20often%20highly%20localized%20and%20personalized%2C%20limiting%20the%20effectiveness%20of%20general-purpose%20LLMs.%20To%20address%20this%20challenge%2C%20we%20present%20RAG-VR%2C%20the%20first%203D%20question-answering%20system%20for%20VR%20that%20incorporates%20retrieval-augmented%20generation%20%28RAG%29%2C%20which%20augments%20an%20LLM%20with%20external%20knowledge%20retrieved%20from%20a%20localized%20knowledge%20database%20to%20improve%20the%20answer%20quality.%20RAG-VR%20includes%20a%20pipeline%20for%20extracting%20comprehensive%20knowledge%20about%20virtual%20environments%20and%20user%20conditions%20for%20accurate%20answer%20generation.%20To%20ensure%20efficient%20retrieval%2C%20RAG-VR%20offloads%20the%20retrieval%20process%20to%20a%20nearby%20edge%20server%20and%20uses%20only%20essential%20information%20during%20retrieval.%20Moreover%2C%20we%20train%20the%20retriever%20to%20effectively%20distinguish%20among%20relevant%2C%20irrelevant%2C%20and%20hard-to-differentiate%20information%20in%20relation%20to%20questions.%20RAG-VR%20improves%20answer%20accuracy%20by%2017.9%22%2C%22date%22%3A%222025%22%2C%22language%22%3A%22%22%2C%22DOI%22%3A%22%22%2C%22ISSN%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fwww.semanticscholar.org%5C%2Fpaper%5C%2Faf54e8314d03df54d1e1857096b053692e325cbc%22%2C%22collections%22%3A%5B%22D2WGRHRC%22%2C%22YN3P4HXF%22%5D%2C%22dateModified%22%3A%222025-05-01T14%3A20%3A00Z%22%7D%7D%2C%7B%22key%22%3A%226YJQENBX%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Izquierdo-Domenech%20et%20al.%22%2C%22parsedDate%22%3A%222024%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BIzquierdo-Domenech%2C%20J.%2C%20Linares-Pellicer%2C%20J.%2C%20%26amp%3B%20Ferri-Molla%2C%20I.%20%282024%29.%20Virtual%20Reality%20and%20Language%20Models%2C%20a%20New%20Frontier%20in%20Learning.%20%26lt%3Bi%26gt%3BInternational%20Journal%20of%20Interactive%20Multimedia%20and%20Artificial%20Intelligence%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3B8%26lt%3B%5C%2Fi%26gt%3B%285%29%2C%2046.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.9781%5C%2Fijimai.2024.02.007%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.9781%5C%2Fijimai.2024.02.007%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Virtual%20Reality%20and%20Language%20Models%2C%20a%20New%20Frontier%20in%20Learning%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Juan%22%2C%22lastName%22%3A%22Izquierdo-Domenech%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jordi%22%2C%22lastName%22%3A%22Linares-Pellicer%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Isabel%22%2C%22lastName%22%3A%22Ferri-Molla%22%7D%5D%2C%22abstractNote%22%3A%22%22%2C%22date%22%3A%222024%22%2C%22language%22%3A%22en%22%2C%22DOI%22%3A%2210.9781%5C%2Fijimai.2024.02.007%22%2C%22ISSN%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fwww.semanticscholar.org%5C%2Fpaper%5C%2F16119e5236d9bd344e6d2027424e880cc6966454%22%2C%22collections%22%3A%5B%22D2WGRHRC%22%2C%22YN3P4HXF%22%5D%2C%22dateModified%22%3A%222025-05-01T14%3A10%3A09Z%22%7D%7D%2C%7B%22key%22%3A%22NKT7XGAZ%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Lewis%20et%20al.%22%2C%22parsedDate%22%3A%222020-12-06%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BLewis%2C%20P.%2C%20Perez%2C%20E.%2C%20Piktus%2C%20A.%2C%20Petroni%2C%20F.%2C%20Karpukhin%2C%20V.%2C%20Goyal%2C%20N.%2C%20K%26%23xFC%3Bttler%2C%20H.%2C%20Lewis%2C%20M.%2C%20Yih%2C%20W.%2C%20Rockt%26%23xE4%3Bschel%2C%20T.%2C%20Riedel%2C%20S.%2C%20%26amp%3B%20Kiela%2C%20D.%20%282020%29.%20Retrieval-augmented%20generation%20for%20knowledge-intensive%20NLP%20tasks.%20%26lt%3Bi%26gt%3BProceedings%20of%20the%2034th%20International%20Conference%20on%20Neural%20Information%20Processing%20Systems%26lt%3B%5C%2Fi%26gt%3B%2C%209459%26%23x2013%3B9474.%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22conferencePaper%22%2C%22title%22%3A%22Retrieval-augmented%20generation%20for%20knowledge-intensive%20NLP%20tasks%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Patrick%22%2C%22lastName%22%3A%22Lewis%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Ethan%22%2C%22lastName%22%3A%22Perez%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Aleksandra%22%2C%22lastName%22%3A%22Piktus%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Fabio%22%2C%22lastName%22%3A%22Petroni%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Vladimir%22%2C%22lastName%22%3A%22Karpukhin%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Naman%22%2C%22lastName%22%3A%22Goyal%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Heinrich%22%2C%22lastName%22%3A%22K%5Cu00fcttler%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Mike%22%2C%22lastName%22%3A%22Lewis%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Wen-tau%22%2C%22lastName%22%3A%22Yih%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Tim%22%2C%22lastName%22%3A%22Rockt%5Cu00e4schel%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Sebastian%22%2C%22lastName%22%3A%22Riedel%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Douwe%22%2C%22lastName%22%3A%22Kiela%22%7D%5D%2C%22abstractNote%22%3A%22Large%20pre-trained%20language%20models%20have%20been%20shown%20to%20store%20factual%20knowledge%20in%20their%20parameters%2C%20and%20achieve%20state-of-the-art%20results%20when%20fine-tuned%20on%20downstream%20NLP%20tasks.%20However%2C%20their%20ability%20to%20access%20and%20precisely%20manipulate%20knowledge%20is%20still%20limited%2C%20and%20hence%20on%20knowledge-intensive%20tasks%2C%20their%20performance%20lags%20behind%20task-specific%20architectures.%20Additionally%2C%20providing%20provenance%20for%20their%20decisions%20and%20updating%20their%20world%20knowledge%20remain%20open%20research%20problems.%20Pre-trained%20models%20with%20a%20differentiable%20access%20mechanism%20to%20explicit%20non-parametric%20memory%20can%20overcome%20this%20issue%2C%20but%20have%20so%20far%20been%20only%20investigated%20for%20extractive%20downstream%20tasks.%20We%20explore%20a%20general-purpose%20fine-tuning%20recipe%20for%20retrieval-augmented%20generation%20%28RAG%29%20%5Cu2014%20models%20which%20combine%20pre-trained%20parametric%20and%20non-parametric%20memory%20for%20language%20generation.%20We%20introduce%20RAG%20models%20where%20the%20parametric%20memory%20is%20a%20pre-trained%20seq2seq%20model%20and%20the%20non-parametric%20memory%20is%20a%20dense%20vector%20index%20of%20Wikipedia%2C%20accessed%20with%20a%20pre-trained%20neural%20retriever.%20We%20compare%20two%20RAG%20formulations%2C%20one%20which%20conditions%20on%20the%20same%20retrieved%20passages%20across%20the%20whole%20generated%20sequence%2C%20and%20another%20which%20can%20use%20different%20passages%20per%20token.%20We%20fine-tune%20and%20evaluate%20our%20models%20on%20a%20wide%20range%20of%20knowledge-intensive%20NLP%20tasks%20and%20set%20the%20state%20of%20the%20art%20on%20three%20open%20domain%20QA%20tasks%2C%20outperforming%20parametric%20seq2seq%20models%20and%20task-specific%20retrieve-and-extract%20architectures.%20For%20language%20generation%20tasks%2C%20we%20find%20that%20RAG%20models%20generate%20more%20specific%2C%20diverse%20and%20factual%20language%20than%20a%20state-of-the-art%20parametric-only%20seq2seq%20baseline.%22%2C%22date%22%3A%22December%206%2C%202020%22%2C%22proceedingsTitle%22%3A%22Proceedings%20of%20the%2034th%20International%20Conference%20on%20Neural%20Information%20Processing%20Systems%22%2C%22conferenceName%22%3A%22%22%2C%22language%22%3A%22%22%2C%22DOI%22%3A%22%22%2C%22ISBN%22%3A%22978-1-7138-2954-6%22%2C%22url%22%3A%22%22%2C%22collections%22%3A%5B%22D2WGRHRC%22%5D%2C%22dateModified%22%3A%222025-05-01T09%3A44%3A23Z%22%7D%7D%2C%7B%22key%22%3A%22VN3R262U%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Marquardt%20et%20al.%22%2C%22parsedDate%22%3A%222025-03-01%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BMarquardt%2C%20A.%2C%20Golchinfar%2C%20D.%2C%20%26amp%3B%20Vaziri%2C%20D.%20%282025%29.%20%26lt%3Bi%26gt%3BRAGatar%3A%20Enhancing%20LLM-driven%20Avatars%20with%20RAG%20for%20Knowledge-Adaptive%20Conversations%20in%20Virtual%20Reality%26lt%3B%5C%2Fi%26gt%3B.%201604%26%23x2013%3B1605.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1109%5C%2FVRW66409.2025.00447%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1109%5C%2FVRW66409.2025.00447%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22conferencePaper%22%2C%22title%22%3A%22RAGatar%3A%20Enhancing%20LLM-driven%20Avatars%20with%20RAG%20for%20Knowledge-Adaptive%20Conversations%20in%20Virtual%20Reality%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Alexander%22%2C%22lastName%22%3A%22Marquardt%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22David%22%2C%22lastName%22%3A%22Golchinfar%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Daryoush%22%2C%22lastName%22%3A%22Vaziri%22%7D%5D%2C%22abstractNote%22%3A%22We%20present%20a%20virtual%20reality%20system%20that%20enables%20users%20to%20seamlessly%20switch%20between%20general%20conversations%20and%20domain-specific%20knowledge%20retrieval%20through%20natural%20interactions%20with%20AI-driven%20avatars.%20By%20combining%20MetaHuman%20technology%20with%20self-hosted%20large%20language%20models%20and%20retrieval-augmented%20generation%2C%20our%20system%20demonstrates%20how%20immersive%20AI%20interactions%20can%20enhance%20learning%20and%20training%20applications%20where%20both%20general%20communication%20and%20expert%20knowledge%20are%20required.%22%2C%22date%22%3A%222025%5C%2F03%5C%2F01%22%2C%22proceedingsTitle%22%3A%22%22%2C%22conferenceName%22%3A%22C%22%2C%22language%22%3A%22English%22%2C%22DOI%22%3A%2210.1109%5C%2FVRW66409.2025.00447%22%2C%22ISBN%22%3A%22979-8-3315-1484-6%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fwww.computer.org%5C%2Fcsdl%5C%2Fproceedings-article%5C%2Fvrw%5C%2F2025%5C%2F148400b604%5C%2F26aUZtE5I1q%22%2C%22collections%22%3A%5B%224B93DUF2%22%5D%2C%22dateModified%22%3A%222025-04-30T14%3A57%3A53Z%22%7D%7D%2C%7B%22key%22%3A%22WJINJTK6%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22N%5Cu00e9meth%20et%20al.%22%2C%22parsedDate%22%3A%222024%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BN%26%23xE9%3Bmeth%2C%20R.%2C%20T%26%23xE1%3Btrai%2C%20A.%2C%20Szab%26%23xF3%3B%2C%20M.%2C%20%26amp%3B%20Tam%26%23xE1%3Bsi%2C%20%26%23xC1%3B.%20%282024%29.%20Using%20a%20RAG-enhanced%20large%20language%20model%26%23xA0%3B%20in%20a%20virtual%20teaching%20assistant%20role%3A%20Experiences%20from%20a%20pilot%20project%20in%20statistics%20education.%20%26lt%3Bi%26gt%3BHungarian%20Statistical%20Review%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3B7%26lt%3B%5C%2Fi%26gt%3B%282%29%2C%203%26%23x2013%3B27.%20Crossref.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.35618%5C%2Fhsr2024.02.en003%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.35618%5C%2Fhsr2024.02.en003%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Using%20a%20RAG-enhanced%20large%20language%20model%20%20in%20a%20virtual%20teaching%20assistant%20role%3A%20Experiences%20from%20a%20pilot%20project%20in%20statistics%20education%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Ren%5Cu00e1ta%22%2C%22lastName%22%3A%22N%5Cu00e9meth%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Annam%5Cu00e1ria%22%2C%22lastName%22%3A%22T%5Cu00e1trai%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Mikl%5Cu00f3s%22%2C%22lastName%22%3A%22Szab%5Cu00f3%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22%5Cu00c1rp%5Cu00e1d%22%2C%22lastName%22%3A%22Tam%5Cu00e1si%22%7D%5D%2C%22abstractNote%22%3A%22The%20role%20of%20artificial%20intelligence%20%28AI%29%20in%20education%20is%20expected%20to%20grow%2C%20but%20how%20it%20transforms%20teaching%5Cn%5Cnand%20learning%20remains%20unclear.%20This%20study%20explores%20the%20use%20of%20an%20AI%20tutor%20that%20is%20similar%20to%20ChatGPT%5Cn%5Cnenhanced%20with%20retrieval-augmented%20generation%20%28RAG%29%2C%20in%20a%20pilot%20project%20at%20the%20Faculty%20of%20Social%5Cn%5CnSciences%20of%20E%5Cu00f6tv%5Cu00f6s%20Lor%5Cu00e1nd%20University%20in%20Budapest%2C%20Hungary.%20The%20tutor%20provided%20a%20searchable%5Cn%5Cnknowledge%20base%20for%20students%20preparing%20for%20admission%20to%20the%20MSc%20in%20Survey%20Statistics%20and%20Data%5Cn%5CnAnalytics.%20Instructor%20feedback%20highlighted%20the%20tutor%5Cu2019s%20ability%20to%20deliver%20accurate%2C%20textbook-based%5Cn%5Cnresponses%2C%20but%20noted%20limitations%20in%20addressing%20real-world%5Cncomplexities.%20Student%20feedback%2C%20which%20was%5Cn%5Cngathered%20through%20focus%20groups%20and%20surveys%2C%20showed%20high%20satisfaction%20and%20many%20used%20the%20tool%20for%5Cn%5Cnactive%20learning%20such%20as%20comparing%20concepts%20and%20organising%20material.%20Students%20had%20the%20flexibility%20to%5Cn%5Cnadapt%20the%20tutor%20to%20their%20own%20learning%20strategy%2C%20and%20they%20also%20noted%20the%20importance%20of%20the%20tutor%20as%20a%5Cn%5Cntime-saving%20supplement%20rather%20than%20a%20replacement%20for%20comprehensive%20study.%20Approximately%2015%25%20of%5Cn%5Cnstudent%20queries%20demonstrated%20critical%20thinking%2C%20where%20students%20used%20the%20AI%20tutor%20to%20confirm%20their%20own%5Cn%5Cninterpretations.%20Similarly%2C%20around%2015%25%20showed%20active%20learning%2C%20seeking%20explanations%20and%5Cn%5Cncomparisons%20or%20generated%20study%20guides%2C%20while%20nearly%2030%25%20engaged%5Cndirectly%20with%20course%20material%2C%5Cn%5Cnreferencing%20specific%20concepts%20and%20theories%20from%20their%20readings.%5CnInstructor%20evaluation%20revealed%20that%5Cn%5Cn76%25%20of%20the%20AI%20tutor%5Cu2019s%20responses%20were%20fully%20correct%2C%2017%25%20mostly%20correct%20and%20only%206%25%20were%20misleading.%5Cn%5CnThe%20findings%20suggest%20that%20RAG%20models%20hold%20promise%20for%20enhancing%5Cnlearning%20by%20offering%20reliable%2C%5Cn%5Cninteractive%20and%20efficient%20support%20for%20students%20and%20educators.%22%2C%22date%22%3A%222024%22%2C%22language%22%3A%22%22%2C%22DOI%22%3A%2210.35618%5C%2Fhsr2024.02.en003%22%2C%22ISSN%22%3A%222630-9130%22%2C%22url%22%3A%22http%3A%5C%2F%5C%2Fdx.doi.org%5C%2F10.35618%5C%2Fhsr2024.02.en003%22%2C%22collections%22%3A%5B%224B93DUF2%22%5D%2C%22dateModified%22%3A%222025-04-30T11%3A40%3A55Z%22%7D%7D%2C%7B%22key%22%3A%22VAW4G3BU%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Prasongpongchai%20et%20al.%22%2C%22parsedDate%22%3A%222024%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BPrasongpongchai%2C%20T.%2C%20Pataranutaporn%2C%20P.%2C%20Kanapornchai%2C%20C.%2C%20Lapapirojn%2C%20A.%2C%20Ouppaphan%2C%20P.%2C%20Winson%2C%20K.%2C%20Lertsutthiwong%2C%20M.%2C%20%26amp%3B%20Maes%2C%20P.%20%282024%29.%20Interactive%20AI-Generated%20Virtual%20Instructors%20Enhance%20Learning%20Motivation%20and%20Engagement%20in%20Financial%20Education.%20In%20A.%20M.%20Olney%2C%20I.-A.%20Chounta%2C%20Z.%20Liu%2C%20O.%20C.%20Santos%2C%20%26amp%3B%20I.%20I.%20Bittencourt%20%28Eds.%29%2C%20%26lt%3Bi%26gt%3BArtificial%20Intelligence%20in%20Education.%20Posters%20and%20Late%20Breaking%20Results%2C%20Workshops%20and%20Tutorials%2C%20Industry%20and%20Innovation%20Tracks%2C%20Practitioners%2C%20Doctoral%20Consortium%20and%20Blue%20Sky%26lt%3B%5C%2Fi%26gt%3B%20%28Vol.%202151%2C%20pp.%20217%26%23x2013%3B225%29.%20Springer%20Nature%20Switzerland.%20https%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1007%5C%2F978-3-031-64312-5_26%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22bookSection%22%2C%22title%22%3A%22Interactive%20AI-Generated%20Virtual%20Instructors%20Enhance%20Learning%20Motivation%20and%20Engagement%20in%20Financial%20Education%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Andrew%20M.%22%2C%22lastName%22%3A%22Olney%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Irene-Angelica%22%2C%22lastName%22%3A%22Chounta%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Zitao%22%2C%22lastName%22%3A%22Liu%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Olga%20C.%22%2C%22lastName%22%3A%22Santos%22%7D%2C%7B%22creatorType%22%3A%22editor%22%2C%22firstName%22%3A%22Ig%20Ibert%22%2C%22lastName%22%3A%22Bittencourt%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Thanawit%22%2C%22lastName%22%3A%22Prasongpongchai%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Pat%22%2C%22lastName%22%3A%22Pataranutaporn%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Chonnipa%22%2C%22lastName%22%3A%22Kanapornchai%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Auttasak%22%2C%22lastName%22%3A%22Lapapirojn%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Pichayoot%22%2C%22lastName%22%3A%22Ouppaphan%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Kavin%22%2C%22lastName%22%3A%22Winson%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Monchai%22%2C%22lastName%22%3A%22Lertsutthiwong%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Pattie%22%2C%22lastName%22%3A%22Maes%22%7D%5D%2C%22abstractNote%22%3A%22%22%2C%22bookTitle%22%3A%22Artificial%20Intelligence%20in%20Education.%20Posters%20and%20Late%20Breaking%20Results%2C%20Workshops%20and%20Tutorials%2C%20Industry%20and%20Innovation%20Tracks%2C%20Practitioners%2C%20Doctoral%20Consortium%20and%20Blue%20Sky%22%2C%22date%22%3A%222024%22%2C%22language%22%3A%22en%22%2C%22ISBN%22%3A%22978-3-031-64311-8%20978-3-031-64312-5%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Flink.springer.com%5C%2F10.1007%5C%2F978-3-031-64312-5_26%22%2C%22collections%22%3A%5B%224B93DUF2%22%5D%2C%22dateModified%22%3A%222025-04-24T13%3A18%3A25Z%22%7D%7D%2C%7B%22key%22%3A%22RASV3NUR%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Maslych%20et%20al.%22%2C%22parsedDate%22%3A%222024%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BMaslych%2C%20M.%2C%20Pumarada%2C%20C.%2C%20Ghasemaghaei%2C%20A.%2C%20%26amp%3B%20LaViola%2C%20J.%20J.%20%282024%29.%20Takeaways%20from%20Applying%20LLM%20Capabilities%20to%20Multiple%20Conversational%20Avatars%20in%20a%20VR%20Pilot%20Study.%20%26lt%3Bi%26gt%3BArXiv%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3Babs%5C%2F2501.00168%26lt%3B%5C%2Fi%26gt%3B%2C%20null.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.48550%5C%2FarXiv.2501.00168%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.48550%5C%2FarXiv.2501.00168%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Takeaways%20from%20Applying%20LLM%20Capabilities%20to%20Multiple%20Conversational%20Avatars%20in%20a%20VR%20Pilot%20Study%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Mykola%22%2C%22lastName%22%3A%22Maslych%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Christian%22%2C%22lastName%22%3A%22Pumarada%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Amirpouya%22%2C%22lastName%22%3A%22Ghasemaghaei%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Joseph%20J.%22%2C%22lastName%22%3A%22LaViola%22%7D%5D%2C%22abstractNote%22%3A%22%22%2C%22date%22%3A%222024%22%2C%22language%22%3A%22%22%2C%22DOI%22%3A%2210.48550%5C%2FarXiv.2501.00168%22%2C%22ISSN%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fwww.semanticscholar.org%5C%2Fpaper%5C%2Fed6d8c1d45e84bd0ae6f94c8ab5106dd1a60819f%22%2C%22collections%22%3A%5B%22D2WGRHRC%22%2C%224B93DUF2%22%2C%22YN3P4HXF%22%5D%2C%22dateModified%22%3A%222025-04-24T08%3A36%3A25Z%22%7D%7D%2C%7B%22key%22%3A%22KPPH5FNM%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Zhang%20et%20al.%22%2C%22parsedDate%22%3A%222023-04-19%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BZhang%2C%20R.%2C%20Zou%2C%20D.%2C%20%26amp%3B%20Cheng%2C%20G.%20%282023%29.%20A%20review%20of%20chatbot-assisted%20learning%3A%20pedagogical%20approaches%2C%20implementations%2C%20factors%20leading%20to%20effectiveness%2C%20theories%2C%20and%20future%20directions.%20%26lt%3Bi%26gt%3BInteractive%20Learning%20Environments%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3B32%26lt%3B%5C%2Fi%26gt%3B%288%29%2C%204529%26%23x2013%3B4557.%20Crossref.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1080%5C%2F10494820.2023.2202704%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1080%5C%2F10494820.2023.2202704%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22A%20review%20of%20chatbot-assisted%20learning%3A%20pedagogical%20approaches%2C%20implementations%2C%20factors%20leading%20to%20effectiveness%2C%20theories%2C%20and%20future%20directions%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Ruofei%22%2C%22lastName%22%3A%22Zhang%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Di%22%2C%22lastName%22%3A%22Zou%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Gary%22%2C%22lastName%22%3A%22Cheng%22%7D%5D%2C%22abstractNote%22%3A%22The%20chatbot%20has%20been%20increasingly%20applied%20and%20investigated%20in%20education%2C%20along%20with%20many%20review%20studies%20from%20different%20aspects.%20However%2C%20few%20reviews%20have%20been%20conducted%20on%20chatbot-assisted%20learning%20from%20the%20pedagogical%20and%20implementational%20aspects%2C%20which%20may%20provide%20implications%20for%20future%20application%20and%20investigation%20of%20educational%20chatbots.%20To%20fill%20in%20the%20gaps%2C%20we%20reviewed%20relevant%20studies%20from%20the%20pedagogical%20and%20implementational%20aspects.%20Forty-six%20articles%20from%20Web%20of%20Science%20and%20Scopus%20databases%20were%20screened%20by%20predefined%20criteria%20and%20analysed%20step%20by%20step%20following%20the%20PRISMA%20framework.%20The%20finding%20showed%20diversified%20learning%20activities%20%28i.e.%20exercise%2C%20instructions%2C%20role-playing%20activities%2C%20collaborative%20product%20design%2C%20independent%20writing%2C%20storytelling%5C%2Fbook-reading%2C%20digital%20gameplay%2C%20and%20open-ended%20debates%29%20that%20chatbots%20could%20support%20through%20presenting%20knowledge%2C%20facilitating%20practices%2C%20supervising%20and%20guiding%20learning%20activities%2C%20and%20providing%20emotional%20support.%20Chabot-assisted%20learning%20was%20applied%20in%2014%20disciplines%2C%20mostly%20in-class%20for%20one%20session%2C%20and%20had%20overall%20positive%20outcomes%20from%20academic%20and%20affective%20aspects.%20Based%20on%20the%20review%20results%2C%20we%20proposed%20a%20RAISE%20model%20of%20effective%5Cnchatbot-assisted%20learning%3A%20Repetitiveness%2C%20Authenticity%2C%5CnInteractivity%2C%20Student-centredness%2C%20and%20Enjoyment.%20We%20identified%20eght%20theories%20that%20might%20be%20useful%20in%20analysing%20and%20supporting%5Cnchatbot-assisted%20learning%3A%20constructivist%20theories%2C%5Cnsituated%5C%2Fcontextualised%20learning%20theories%2C%20cognitive%20theories%20of%5Cnmultimedia%20learning%2C%20self-regulated%20learning%20theories%2C%20output%5Cnhypotheses%2C%20flow%20theory%2C%20collaborative%20learning%20theories%20and%5Cnmotivation%20theories.%20Future%20studies%20on%20chatbot-assisted%20learning%20may%20be%20conducted%20on%20the%20use%20of%20theoretical%20frameworks%2C%20the%20application%20of%20various%20technological-pedagogical%20approaches%20and%20learning%20activities%2C%20and%20the%20long-term%2C%20out-of-class%20implementations%20in%20new%20areas.%22%2C%22date%22%3A%222023-4-19%22%2C%22language%22%3A%22en%22%2C%22DOI%22%3A%2210.1080%5C%2F10494820.2023.2202704%22%2C%22ISSN%22%3A%221049-4820%22%2C%22url%22%3A%22http%3A%5C%2F%5C%2Fdx.doi.org%5C%2F10.1080%5C%2F10494820.2023.2202704%22%2C%22collections%22%3A%5B%224B93DUF2%22%5D%2C%22dateModified%22%3A%222025-04-24T08%3A36%3A16Z%22%7D%7D%2C%7B%22key%22%3A%224IC89ZTV%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Thway%20et%20al.%22%2C%22parsedDate%22%3A%222024%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BThway%2C%20M.%2C%20Recatala-Gomez%2C%20J.%2C%20Lim%2C%20F.%20S.%2C%20Hippalgaonkar%2C%20K.%2C%20%26amp%3B%20Ng%2C%20L.%20W.%20T.%20%282024%29.%20Battling%20Botpoop%20using%20GenAI%20for%20Higher%20Education%3A%20A%20Study%20of%20a%20Retrieval%20Augmented%20Generation%20Chatbots%20Impact%20on%20Learning.%20%26lt%3Bi%26gt%3BArXiv%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3Babs%5C%2F2406.07796%26lt%3B%5C%2Fi%26gt%3B%2C%20null.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.48550%5C%2FarXiv.2406.07796%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.48550%5C%2FarXiv.2406.07796%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Battling%20Botpoop%20using%20GenAI%20for%20Higher%20Education%3A%20A%20Study%20of%20a%20Retrieval%20Augmented%20Generation%20Chatbots%20Impact%20on%20Learning%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Maung%22%2C%22lastName%22%3A%22Thway%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Jose%22%2C%22lastName%22%3A%22Recatala-Gomez%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Fun%20Siong%22%2C%22lastName%22%3A%22Lim%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Kedar%22%2C%22lastName%22%3A%22Hippalgaonkar%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Leonard%20W.%20T.%22%2C%22lastName%22%3A%22Ng%22%7D%5D%2C%22abstractNote%22%3A%22%22%2C%22date%22%3A%222024%22%2C%22language%22%3A%22%22%2C%22DOI%22%3A%2210.48550%5C%2FarXiv.2406.07796%22%2C%22ISSN%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fwww.semanticscholar.org%5C%2Fpaper%5C%2F79a6d1bd9dfad6c858369015cae56cf1e1ad3f9d%22%2C%22collections%22%3A%5B%22D2WGRHRC%22%2C%224B93DUF2%22%5D%2C%22dateModified%22%3A%222025-04-23T15%3A43%3A07Z%22%7D%7D%2C%7B%22key%22%3A%22MGRNGC8L%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22edenai%22%2C%22parsedDate%22%3A%222025-03-17%22%2C%22numChildren%22%3A0%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3Bedenai.%20%282025%29.%20%26lt%3Bi%26gt%3Bedenai%5C%2Funity-plugin%26lt%3B%5C%2Fi%26gt%3B.%20Eden%20AI.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-ItemURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fgithub.com%5C%2Fedenai%5C%2Funity-plugin%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fgithub.com%5C%2Fedenai%5C%2Funity-plugin%26lt%3B%5C%2Fa%26gt%3B%20%28Original%20work%20published%202023%29%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22computerProgram%22%2C%22title%22%3A%22edenai%5C%2Funity-plugin%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22programmer%22%2C%22name%22%3A%22edenai%22%7D%5D%2C%22abstractNote%22%3A%22The%20Unity%20EdenAI%20Plugin%20simplifies%20integrating%20AI%20tasks%20like%20text-to-speech%2C%20chatbots%20and%20other%20generative%20AI%20into%20Unity%20applications%20using%20the%20EdenAI%20API.%22%2C%22versionNumber%22%3A%22%22%2C%22date%22%3A%222025-03-17T18%3A42%3A15Z%22%2C%22system%22%3A%22%22%2C%22company%22%3A%22Eden%20AI%22%2C%22programmingLanguage%22%3A%22C%23%22%2C%22ISBN%22%3A%22%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fgithub.com%5C%2Fedenai%5C%2Funity-plugin%22%2C%22collections%22%3A%5B%227ME8Z5N3%22%5D%2C%22dateModified%22%3A%222025-04-01T17%3A37%3A43Z%22%7D%7D%2C%7B%22key%22%3A%226DRDGIID%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Ortega%5Cu2010Ochoa%20et%20al.%22%2C%22parsedDate%22%3A%222023-12-06%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BOrtega%26%23x2010%3BOchoa%2C%20E.%2C%20Arguedas%2C%20M.%2C%20%26amp%3B%20Daradoumis%2C%20T.%20%282023%29.%20Empathic%20pedagogical%20conversational%20agents%3A%20A%20systematic%20literature%20review.%20%26lt%3Bi%26gt%3BBritish%20Journal%20of%20Educational%20Technology%26lt%3B%5C%2Fi%26gt%3B%2C%20%26lt%3Bi%26gt%3B55%26lt%3B%5C%2Fi%26gt%3B%283%29%2C%20886%26%23x2013%3B909.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-DOIURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1111%5C%2Fbjet.13413%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdoi.org%5C%2F10.1111%5C%2Fbjet.13413%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22journalArticle%22%2C%22title%22%3A%22Empathic%20pedagogical%20conversational%20agents%3A%20A%20systematic%20literature%20review%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Elvis%22%2C%22lastName%22%3A%22Ortega%5Cu2010Ochoa%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Marta%22%2C%22lastName%22%3A%22Arguedas%22%7D%2C%7B%22creatorType%22%3A%22author%22%2C%22firstName%22%3A%22Thanasis%22%2C%22lastName%22%3A%22Daradoumis%22%7D%5D%2C%22abstractNote%22%3A%22Artificial%20intelligence%20%28AI%29%20and%20natural%20language%20processing%20technologies%20have%20fuelled%20the%20growth%20of%20Pedagogical%20Conversational%20Agents%20%28PCAs%29%20with%20empathic%20conversational%20capabilities.%20However%2C%20no%20systematic%20literature%20review%20has%20explored%20the%20intersection%20between%20conversational%20agents%2C%20education%20and%20emotion.%20Therefore%2C%20this%20study%20aimed%20to%20outline%20the%20key%20aspects%20of%20designing%2C%20implementing%20and%20evaluating%20these%20agents.%20The%20data%20sources%20were%20empirical%20studies%2C%20including%20peer-reviewed%20conference%20papers%20and%20journal%20articles%2C%20and%20the%20most%20recent%20publications%2C%20from%20the%20ACM%20Digital%20Library%2C%20IEEE%20Xplore%2C%20ProQuest%2C%20ScienceDirect%2C%20Scopus%2C%20SpringerLink%2C%20Taylor%20%26amp%3B%20Francis%20Online%2C%20Web%20of%20Science%20and%20Wiley%20Online%20Library.%20The%20remaining%20papers%20underwent%20a%20rigorous%20quality%20assessment.%20A%20filter%20study%20meeting%20the%20objective%20was%20based%20on%20keywords.%20Comparative%20analysis%20and%20synthesis%20of%20results%20were%20used%20to%20handle%20data%20and%20combine%20study%20outcomes.%20Out%20of%201162%20search%20results%2C%2013%20studies%20were%20selected.%20The%20results%20indicate%20that%20agents%20promote%20dialogic%20learning%2C%20proficiency%20in%20knowledge%20domains%2C%20personalized%20feedback%20and%20empathic%20abilities%20as%20essential%20design%20principles.%20Most%20implementations%20employ%20a%20quantitative%20approach%2C%20and%20two%20variables%20are%20used%20for%20evaluation.%20Feedback%20types%20play%20a%20vital%20role%20in%20achieving%20positive%20results%20in%20learning%20performance%20and%20student%20perceptions.%20The%20main%20limitations%20and%20gaps%20are%20the%20time%20range%20for%20literature%20selection%2C%20the%20level%20of%20integration%20of%20the%20empathic%20field%20and%20the%20lack%20of%20a%20detailed%20development%20stage%20report.%20Moreover%2C%20future%20directions%20are%20the%20ethical%20implications%20of%20agents%20operating%20beyond%20scheduled%20learning%20times%20and%20the%20adoption%20of%20Responsible%20AI%20principles.%20In%20conclusion%2C%20this%20review%20provides%20a%20comprehensive%20framework%20of%20empathic%20PCAs%2C%20mostly%20in%20their%20evaluation.%20The%20systematic%20review%20registration%20number%20is%20osf.io%5C%2F3xk6a.Practitioner%20notesWhat%20is%20already%20known%20about%20this%20topic%20Emotions%20play%20a%20pivotal%20role%20in%20shaping%20the%20interaction%20process%2C%20making%20it%20essential%20to%20consider%20them%20when%20designing%20methodological%20strategies%20or%20learning%20tools.%20Empathic%20Pedagogical%20Conversational%20Agents%20%28PCAs%29%20have%20emerged%20as%20a%20crucial%20approach%20for%20enhancing%20and%20personalizing%20the%20learning%20experience%20%2824%5C%2F7%29%20for%20pupils%20and%20supporting%20human%20teachers%20in%20their%20teaching%20process.%20Despite%20the%20creation%20of%20numerous%20empathic%20PCAs%2C%20there%20is%20a%20scarcity%20of%20Systematic%20Literature%20Reviews%20%28SLRs%29%20on%20their%20application%20in%20the%20educational%20field%2C%20particularly%20concerning%20the%20integration%20of%20emotional%20abilities%20in%20combination%20with%20the%20competencies%20of%20each%20subject.%20What%20this%20paper%20adds%20It%20offers%20new%20insights%20into%20the%20design%20principles%20underlying%20the%20integration%20of%20the%20empathic%20field.%20It%20reviews%20different%20approaches%20for%20incorporating%20students%26%23039%3B%20prior%20knowledge%20in%20real%20time.%20It%20provides%20a%20comprehensive%20and%20up-to-date%20overview%20of%20the%20research%20designs%20used%20for%20implementation%2C%20including%20quantitative%2C%20qualitative%20and%20mixed%20methods.%20It%20examines%20the%20factors%20that%20influence%20the%20effectiveness%20of%20empathic%20PCA%20in%20teaching%20and%20learning.%20It%20evaluates%20the%20types%20of%20feedback%20that%20enhance%20the%20impact%20of%20the%20empathic%20field%20on%20learning%20outcomes.%20Implications%20for%20practice%20and%5C%2For%20policy%20It%20is%20crucial%20to%20grasp%20the%20topics%20that%20this%20paper%20introduces%20in%20order%20to%20effectively%20integrate%20new%20learning%20tools%20into%20any%20context.%20Techno-pedagogical%20designers%20seeking%20to%20gain%20insights%20into%20empathic%20PCAs%20will%20find%20immense%20value%20in%20this%20SLR%2C%20as%20it%20comprehensively%20covers%20each%20stage%20of%20the%20process.%20For%20future%20research%20endeavours%2C%20this%20study%20offers%20a%20wealth%20of%20ideas%20to%20draw%20upon%2C%20enabling%20researchers%20to%20address%20the%20challenges%20outlined%20and%20explore%20new%20avenues%20of%20investigation.%22%2C%22date%22%3A%222023-12-6%22%2C%22language%22%3A%22en%22%2C%22DOI%22%3A%2210.1111%5C%2Fbjet.13413%22%2C%22ISSN%22%3A%221467-8535%22%2C%22url%22%3A%22%22%2C%22collections%22%3A%5B%22JIGNPJL6%22%2C%22XNETGP47%22%5D%2C%22dateModified%22%3A%222025-04-01T17%3A11%3A49Z%22%7D%7D%2C%7B%22key%22%3A%22ZCZGX2AY%22%2C%22library%22%3A%7B%22id%22%3A5189002%7D%2C%22meta%22%3A%7B%22creatorSummary%22%3A%22Meta%22%2C%22parsedDate%22%3A%222024-08-19%22%2C%22numChildren%22%3A1%7D%2C%22bib%22%3A%22%26lt%3Bdiv%20class%3D%26quot%3Bcsl-bib-body%26quot%3B%20style%3D%26quot%3Bline-height%3A%202%3B%20padding-left%3A%201em%3B%20text-indent%3A-1em%3B%26quot%3B%26gt%3B%5Cn%20%20%26lt%3Bdiv%20class%3D%26quot%3Bcsl-entry%26quot%3B%26gt%3BMeta.%20%282024%2C%20August%2019%29.%20%26lt%3Bi%26gt%3BVoice%20SDK%20Overview%26lt%3B%5C%2Fi%26gt%3B.%20Meta%20Horizon%20OS%20Developers.%20%26lt%3Ba%20class%3D%26%23039%3Bzp-ItemURL%26%23039%3B%20href%3D%26%23039%3Bhttps%3A%5C%2F%5C%2Fdevelopers.meta.com%5C%2Fhorizon%5C%2Fdocumentation%5C%2Funity%5C%2Fvoice-sdk-overview%5C%2F%26%23039%3B%26gt%3Bhttps%3A%5C%2F%5C%2Fdevelopers.meta.com%5C%2Fhorizon%5C%2Fdocumentation%5C%2Funity%5C%2Fvoice-sdk-overview%5C%2F%26lt%3B%5C%2Fa%26gt%3B%26lt%3B%5C%2Fdiv%26gt%3B%5Cn%26lt%3B%5C%2Fdiv%26gt%3B%22%2C%22data%22%3A%7B%22itemType%22%3A%22webpage%22%2C%22title%22%3A%22Voice%20SDK%20Overview%22%2C%22creators%22%3A%5B%7B%22creatorType%22%3A%22author%22%2C%22name%22%3A%22Meta%22%7D%5D%2C%22abstractNote%22%3A%22This%20section%20provides%20an%20overview%20of%20the%20Voice%20SDK.%22%2C%22date%22%3A%2219%5C%2F08%5C%2F2024%22%2C%22url%22%3A%22https%3A%5C%2F%5C%2Fdevelopers.meta.com%5C%2Fhorizon%5C%2Fdocumentation%5C%2Funity%5C%2Fvoice-sdk-overview%5C%2F%22%2C%22language%22%3A%22en%22%2C%22collections%22%3A%5B%227ME8Z5N3%22%5D%2C%22dateModified%22%3A%222025-04-01T16%3A35%3A32Z%22%7D%7D%5D%7D
Ding, S., & Chen, Y. (2025). RAG-VR: Leveraging retrieval-augmented generation for 3D question answering in VR environments. https://www.semanticscholar.org/paper/af54e8314d03df54d1e1857096b053692e325cbc
Izquierdo-Domenech, J., Linares-Pellicer, J., & Ferri-Molla, I. (2024). Virtual Reality and Language Models, a New Frontier in Learning. International Journal of Interactive Multimedia and Artificial Intelligence, 8(5), 46. https://doi.org/10.9781/ijimai.2024.02.007
Lewis, P., Perez, E., Piktus, A., Petroni, F., Karpukhin, V., Goyal, N., Küttler, H., Lewis, M., Yih, W., Rocktäschel, T., Riedel, S., & Kiela, D. (2020). Retrieval-augmented generation for knowledge-intensive NLP tasks. Proceedings of the 34th International Conference on Neural Information Processing Systems, 9459–9474.
Marquardt, A., Golchinfar, D., & Vaziri, D. (2025). RAGatar: Enhancing LLM-driven Avatars with RAG for Knowledge-Adaptive Conversations in Virtual Reality. 1604–1605. https://doi.org/10.1109/VRW66409.2025.00447
Németh, R., Tátrai, A., Szabó, M., & Tamási, Á. (2024). Using a RAG-enhanced large language model in a virtual teaching assistant role: Experiences from a pilot project in statistics education. Hungarian Statistical Review, 7(2), 3–27. Crossref. https://doi.org/10.35618/hsr2024.02.en003
Prasongpongchai, T., Pataranutaporn, P., Kanapornchai, C., Lapapirojn, A., Ouppaphan, P., Winson, K., Lertsutthiwong, M., & Maes, P. (2024). Interactive AI-Generated Virtual Instructors Enhance Learning Motivation and Engagement in Financial Education. In A. M. Olney, I.-A. Chounta, Z. Liu, O. C. Santos, & I. I. Bittencourt (Eds.), Artificial Intelligence in Education. Posters and Late Breaking Results, Workshops and Tutorials, Industry and Innovation Tracks, Practitioners, Doctoral Consortium and Blue Sky (Vol. 2151, pp. 217–225). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-64312-5_26
Maslych, M., Pumarada, C., Ghasemaghaei, A., & LaViola, J. J. (2024). Takeaways from Applying LLM Capabilities to Multiple Conversational Avatars in a VR Pilot Study. ArXiv, abs/2501.00168, null. https://doi.org/10.48550/arXiv.2501.00168
Zhang, R., Zou, D., & Cheng, G. (2023). A review of chatbot-assisted learning: pedagogical approaches, implementations, factors leading to effectiveness, theories, and future directions. Interactive Learning Environments, 32(8), 4529–4557. Crossref. https://doi.org/10.1080/10494820.2023.2202704
Thway, M., Recatala-Gomez, J., Lim, F. S., Hippalgaonkar, K., & Ng, L. W. T. (2024). Battling Botpoop using GenAI for Higher Education: A Study of a Retrieval Augmented Generation Chatbots Impact on Learning. ArXiv, abs/2406.07796, null. https://doi.org/10.48550/arXiv.2406.07796
edenai. (2025). edenai/unity-plugin. Eden AI. https://github.com/edenai/unity-plugin (Original work published 2023)
Ortega‐Ochoa, E., Arguedas, M., & Daradoumis, T. (2023). Empathic pedagogical conversational agents: A systematic literature review. British Journal of Educational Technology, 55(3), 886–909. https://doi.org/10.1111/bjet.13413
Meta. (2024, August 19). Voice SDK Overview. Meta Horizon OS Developers. https://developers.meta.com/horizon/documentation/unity/voice-sdk-overview/
Diagram content is licensed under a Creative Commons Attribution-ShareAlike 4.0 International license.