https://arxiv.org/pdf/2402.03578 appears to be quite interesting for my work in the #w3c community group for #voiceinteraction and the Open Voice Network #ovon:
Multi-agent systems boost LLM capabilities via agent collaboration. Challenges include task allocation, reasoning, context & memory management. Explores applications in blockchain systems. #LLMs #MultiAgentSystems #AI #Blockchain
#voiceinteraction
I just added a first draft for #security and #privacy in the context of #smartassistants to the Architecture and Potential for Standardization standardization draft of our #w3c community group for #voiceinteraction.
Any suggestions for improvement are welcome.
NotebookLM expands with Plus tier and interactive audio features: Google introduces premium features and voice interaction for its AI research assistant, targeting enterprise users. https://ppc.land/notebooklm-expands-with-plus-tier-and-interactive-audio-features/?utm_source=dlvr.it&utm_medium=mastodon #NotebookLM #ArtificialIntelligence #VoiceInteraction #AIResearch #EnterpriseSolutions
I made some changes to https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paArchitecture/paArchitecture-1-3.htm of our #w3c #voiceinteraction community group.
Please review and let me know in case something should be described better
1 Introduction
- Added a paragraph to detail how IPA is meant and used in the document
2.1.2 Use Case
- Removed TODOs for Interoperability
2.1.3 In-Vehicle Chatbot
Added a hint on LLMs and GenAI
3.3.1.2 Provider Registry
-Added some hint about discovery to the provider registry component
- Added hint to OVON assistant manifest
I am currently checking again the Open Voice Network #OVON https://openvoicenetwork.org/ and how this can be combined with my work for the #W3C #voiceinteraction community group.
Started to join their meetings.
Changes for https://github.com/w3c/voiceinteraction/issues/52 Improve description of Generative AI elements after the review in the last session on 2024-10-23 of the #w3c #voiceinteraction community group
section 3
- added a note about location of components
section 3.2
- added differences between Trad #NLU and Gen AI system
section 3.2.5.2, 3.3.1
- Added #agenticapplications
section 3.2.8.2, 3.2.8.4
- Spearation of usages in Trad NLU and Gen AI
Figure 3, 4
- gray out #LLM
I completed my work on https://github.com/w3c/voiceinteraction/issues/52 Improve description of Generative AI elements from our #w3c community group for #voiceinteraction. Maybe you still see some room for improvement? Be invited to check and comment.
Changes are available at https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paArchitecture/paArchitecture-1-3.htm
#smartassistants #conversationalai #generativeai #intelligentpersonalassistants
Robin Knote based his PhD also on our efforts to standardize intelligent personal assistants in the #w3c #voiceinteraction community group.
German only: https://kobra.uni-kassel.de/handle/123456789/15444
In the #w3c #voiceinteraction community group we are just discussing if intelligent personal assistants can be called personal if all data is shared with others.
Any opinions on that?
#digitalassistant #voicecommunication #conversationalai #privacy #trust
OpenAI ChatGPT Voice Rolled Out Plus Users, Check How to use it on Mobile.
See here - https://techchilli.com/artificial-intelligence/openai-chatgpt-voice-rolled-out-for-free/
#ChatGPTVoice #OpenAI #AIInnovation #GregBrockman #VoiceAI #AIChat #MobileTech #TechTrends #VoiceInteraction #FutureOfAI #AIRevolution
Discovered a #w3c community group working on #knowledgegraph https://www.w3.org/community/kg-construct/
They focus on the redesign of the current w3c recommendation for generating RDF graphs from RDB databases (R2RML: https://www.w3.org/TR/r2rml/).
I am curious to learn if and how their work could be useful for our efforts for a standarad for intelligent personal
assistants, where knowledge graphs are one aspect that we are considering.
See also our current draft report
https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paArchitecture/paArchitecture-1-3.htm#finalarchitecture
I made some updates to improve the description of #GenerativeAI components.
Current draft is at https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paArchitecture/paArchitecture-1-3.htm
I was under the impression that dialogs with their description how to cope with input and decide upon the output would only be relevant for traditional #NLU based systems. However, in the #LLM post processing we may also be in the need to make use of, e.g., error dialogs if a call to the LLM could not be established. I still need to think about it.
I'm almost done with the walk through of the demo code for our framework of intelligent personal assistants at https://github.com/w3c/voiceinteraction/blob/master/source%2FSOURCE.md. We will continue to talk about it in the next meeting of the #w3c community group for #voiceinteraction.
Some more refactorings in our reference implementation of a framework for intelligent personal assistants that allows for some more flexible chaining of components like this
modalityManager >> inputListener >> ipaService >> providerSelectionService
>> ipaService >> modalityManager;
More details are here https://github.com/w3c/voiceinteraction/blob/master/source/SOURCE.md
#voiceinteraction #foss #opensource #w3c #community #programming #ChatGPT
In yesterday's call of the #w3c community group for #voiceinteraction, we discussed a more visible representation of #LLMs in our reference architecture for #intelligentpersonalassistants #ipa .
A first draft is now available under https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paArchitecture/paArchitecture-1-3.htm , which we will discuss in our next call on 2024-04-10.
The currently followed architecture of our #w3c community group is still too closely bound to traditional #NLU based #voiceinteraction concepts. While it aimed at including #llm with speech, LLM with multimodality, ... it is potentially not visible enough. We may want to also improve here.
Some first thoughts about error handling in intelligent personal assistants from our #w3c #voiceinteraction community group.
Another attempt to define the concept of a session in a #voiceinteraction context from our #w3c community group
I am making slow but constant progress with the reference implementation of our efforts in the #w3c #voiceinteraction community to standardize digital intelligent personal assistants.
So far, the concept sketched at https://w3c.github.io/voiceinteraction/voice%20interaction%20drafts/paInterfaces/paInterfaces.htm seem to be working. Here and there are some smaller hickups.
Nevertheless, it may be good practice for #API designers to verify with an actual implementation that the concepts are actually working.
I just came across Spring AI, who are working on an API to facilitate the creation of AI applications
https://docs.spring.io/spring-ai/reference/getting-started.html
Maybe I should have a look at what they are doing and what can be learned from that for the #w3c #voiceinteraction community group.
Client Info
Server: https://mastodon.social
Version: 2025.04
Repository: https://github.com/cyevgeniy/lmst