[Paper Review] From Google Gemini to OpenAI Q* (Q-Star): A Survey of Reshaping the Generative Artificial Intelligence (AI) Research Landscape

Timothy R. McIntosh, Teo Sušnjak+3 more|arXiv (Cornell University)|Dec 18, 2023

Artificial Intelligence in Healthcare and Education60 citations

TL;DR

This survey analyzes how Mixture of Experts, multimodality, and envisioned AGI reshape the generative AI research landscape, focusing on Gemini and the speculative Q* project and their implications for research taxonomy, ethics, and scholarly communication.

ABSTRACT

This comprehensive survey explored the evolving landscape of generative Artificial Intelligence (AI), with a specific focus on the transformative impacts of Mixture of Experts (MoE), multimodal learning, and the speculated advancements towards Artificial General Intelligence (AGI). It critically examined the current state and future trajectory of generative Artificial Intelligence (AI), exploring how innovations like Google's Gemini and the anticipated OpenAI Q* project are reshaping research priorities and applications across various domains, including an impact analysis on the generative AI research taxonomy. It assessed the computational challenges, scalability, and real-world implications of these technologies while highlighting their potential in driving significant progress in fields like healthcare, finance, and education. It also addressed the emerging academic challenges posed by the proliferation of both AI-themed and AI-generated preprints, examining their impact on the peer-review process and scholarly communication. The study highlighted the importance of incorporating ethical and human-centric methods in AI development, ensuring alignment with societal norms and welfare, and outlined a strategy for future AI research that focuses on a balanced and conscientious use of MoE, multimodality, and AGI in generative AI.

Motivation & Objective

• Examine how Mixture of Experts (MoE), multimodality, and Artificial General Intelligence (AGI) reshape generative AI.
• Assess the implications of Google Gemini and the speculative OpenAI Q* on research priorities and taxonomy.
• Analyze ethical, societal, and technical challenges arising from rapid generative AI development.
• Explore governance, data privacy, and human-centric approaches to AI deployment.
• Highlight academic challenges associated with AI-themed and AI-generated preprints.

Proposed method

• Structured literature survey across IEEE Xplore, Scopus, ACM DL, ScienceDirect, Web of Science, and ProQuest Central.
• Temporal scope from 2017 (Transformer release) to 2023 (manuscript writing).
• Taxonomy development and synthesis of current and emergent trends in generative AI.
• Critical appraisal of MoE, multimodality, and AGI as drivers of the AI research landscape.
• Discussion of ethical, societal, and governance considerations tied to generative AI advances.

Experimental results

Research questions

• RQ1 How do MoE, multimodal AI, and AGI reshape the generative AI research landscape?
• RQ2 What are the implications of Gemini and the speculative Q* project on research taxonomy and priorities?
• RQ3 What ethical, societal, and technical challenges accompany rapid generative AI development?
• RQ4 How do preprints affect peer review and scholarly communication in AI research?

Key findings

• MoE enables scalable, specialized modeling but faces dynamic routing, expert imbalance, and alignment challenges.
• Gemini advances multimodality with broader data types and cross-modal understanding, yet real-world reasoning integration requires further evaluation.
• Speculative Q* envisions integrating LLMs with Q-learning and A* to combine learning, creativity, and structured problem solving beyond current multimodal systems.
• The surge in AI preprints raises concerns about validation, guidance for scholarly communication, and potential dissemination of unvetted results.
• Ethical alignment and human-centric governance remain essential amidst rapid AI advancement and broader societal implications.