【2023 RISC-V Taipei Day】PERSONALIZATION, PRIVACY, DATA PROTECTION, AND LLM ON RISC-V

The AI & automotive industry are undergoing a major transformation as AI is being used to improve safety, efficiency, and comfort in vehicles, as well as to create new personalized experiences for drivers and passengers. At the same time, the Automotive industry has been trying to transform cars into AI enabled Software Defined Vehicles (SDV).
On October 12, RVTA hosted the 2023 RISC-V Taipei Day forum that explored how the RISC-V open source processor architecture can empower complex IC systems and AI computing architectures in the AI era. In this presentation, the speakers discussed the latest trends in the AI & Automotive industries with a focus on personalization, privacy, and data protection. The forum also discussed the advantages of using RISC-V for AI SoC in operating large language models (LLMs).
RISC-V’s open architecture helps AI SoCs run LLMs; offering better security and reliability with more design freedom. As the industry develops AI-powered vehicles, more and more data about their surroundings and occupants will be collected and it is important to ensure that this data is secured and not misused.
RISC-V Taiwan Alliance (RVTA)
RVTA aims to introduce RISC-V’s open architecture to Taiwan through cross-ecosystem cooperation among industries, academia, and research institutes both local and abroad. By enabling AIoT integration in Taiwanese industries from R&D, design, and application processes; they will be able to leverage 5G communication trends and business opportunities and enhance Taiwan’s industrial competitiveness.
--------------------------------------------------------------------------------------------------------------
( 演講內容重點提要 )
支援特定領域用LLM模型AI晶片 搭配End-to-End訓練流程 將是AI設備設計新主流
臺灣發展軟體科技（Skymizer）行銷長暨執行副總魏國章，在「Trends in AI & Automotive: Personalization, Privacy, Date Protection, and LLM on RISC-V」專題演講中表示，由Open AI推出的ChatGPT技術所帶來的AI應用風潮，讓不少廠商開始思考如何將LLM服務放到設備上，而LLM所需要的模型，便成為廠商在開發LLM服務當中要思考的問題，因為除了通用型LLM，特定領域用LLM，甚至到有增加私人資料的個人用LLM，都有不同的資料隱私與模型要處理。也因為LLM的應用範圍相當廣，因此AI客製化晶片未來勢必要支援LLM功能，而RISC-V的開放式架構，可以讓廠商在開發LLM AI晶片上更有自由度。Skymizer可以協助晶片廠設計編程器，讓客製化晶片能和AI深度學習、高速運算整合，提升硬體效能，也能導入In Memory AI運算架構，提升晶片AI運算效率，降低AI運算所需能耗。
針對汽車電子市場，魏國章認為軟體定義汽車（Soft Defined Vehicle，SDV）是電動車發展趨勢，目前市場上以Tesla為主要領導廠商，但是封閉式平台。SDV開源平台包括MIH EVKit SDK、Eclipse SDV WG、SOAFEE、Open AD Kit。目前無人駕駛AI訓練技術有兩個方式，一個是模組加程式碼（Modular + Code），案例是Waymo，另外一個則是End-to-End學習，案例是Open Pilot跟Tesla。而類似的案例也發生在自主機器人身上，Boston Dynamics使用模組加程式碼，而Tesla的Optimus則是使用End-to-End學習。而不同的開發技術，在產品功能進步上其實會有差異，End-to-End學習方式進步速度很明顯快了不少。
針對未來的AI使用者體驗，魏國章認為Domain LLM加上Fine Tuning與AI Agent會是主流。舉例來說，AI電動車的操作會是使用者語音搭配Open Interpreter 再加上EV LLM，如此一來，使用者可以透過語音直接控制電動車，不僅非常直覺，就跟操作霹靂遊俠（Knight Rider）裡面的霹靂車一樣。而電動車的AI晶片，將會是AI晶片加上EV LLM與Agent，然後搭配Safety module，而Skymizer可以協助IC設計公司開發車用AI晶片與對應軟體。
關於台灣RISC-V聯盟
致力於產、學、研三方合作，共同協助將RISC-V開放架構導入台灣，並串聯海內外生態系資源，讓台灣產業從研發、設計到應用，都能具備AIoT整合能力；搭上5G通訊趨勢與商機，進而提升台灣產業競爭力。
www.twiota.org/RISC-V/default....
#AI #Automotive #LLM #RISC-V