International Peer-Reviewed JournalOpen AccessISSN 2456-8880
irejournals@gmail.com+91-7433024337

Home / Current Issue / Paper 1710302

1710302PublishedVol 9 · Issue 2

Mixture-of-Experts-and-Depths: A Hierarchical Dynamic Compute Architecture for Extreme-Scale Efficiency

Kalyan Chakravarthy Kodela

Subject area: Science,Engineering and Technology  ·  Area of research: Artificial Intelligence

Abstract

The pursuit of larger, more capable Large Language Models (LLMs) is fundamentally constrained by the immense computational cost of their training and inference. While the Mixture-of-Experts (MoE) paradigm successfully decouples model parameter count from computational cost by dynamically scaling network width, it neglects the critical dimension of depth, enforcing a uniform and often wasteful computational graph for all tokens. This paper introduces Mixture-of-Experts-and-Depths (MoED), a novel architectural framework that synergistically unifies dynamic width and depth scaling. MoED employs a hierarchical routing mechanism, where a meta-controller at each layer makes a joint decision on both expert selection and a token's subsequent computational path?whether to exit, proceed, or skip ahead. This approach creates a unique, input-adaptive sub-network for every token, optimizing the allocation of compute. The proposed architecture presents a fundamental shift towards more efficient and scalable LLMs, theoretically enabling superior performance and reduced latency while managing the activation memory bottlenecks that plague traditional trillion-parameter models.

Keywords

Large Language Models, Mixture of Experts, Adaptive Computation, Dynamic Networks, Hierarchical Routing, Efficient Inference

How to cite this paper

Kalyan Chakravarthy Kodela "Mixture-of-Experts-and-Depths: A Hierarchical Dynamic Compute Architecture for Extreme-Scale Efficiency" Iconic Research And Engineering Journals Volume 9 Issue 2 2025 Page 987-999
Kalyan Chakravarthy Kodela "Mixture-of-Experts-and-Depths: A Hierarchical Dynamic Compute Architecture for Extreme-Scale Efficiency" Iconic Research And Engineering Journals, vol. 9, no. 2, Aug. 2025
Kalyan Chakravarthy Kodela (2025). Mixture-of-Experts-and-Depths: A Hierarchical Dynamic Compute Architecture for Extreme-Scale Efficiency. Iconic Research And Engineering Journals, 9(2).
Kalyan Chakravarthy Kodela "Mixture-of-Experts-and-Depths: A Hierarchical Dynamic Compute Architecture for Extreme-Scale Efficiency" Iconic Research And Engineering Journals, vol. 9, no. 2, Aug. 2025.
@article{1710302,
      author = {Kalyan Chakravarthy Kodela},
      title = {Mixture-of-Experts-and-Depths: A Hierarchical Dynamic Compute Architecture for Extreme-Scale Efficiency},
      journal = {Iconic Research And Engineering Journals},
      year = {2025},
      volume = {9},
      number = {2},
      pages = {987-999},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1710302.pdf},
      abstract = {The pursuit of larger, more capable Large Language Models (LLMs) is fundamentally constrained by the immense computational cost of their training and inference. While the Mixture-of-Experts (MoE) paradigm successfully decouples model parameter count from computational cost by dynamically scaling network width, it neglects the critical dimension of depth, enforcing a uniform and often wasteful computational graph for all tokens. This paper introduces Mixture-of-Experts-and-Depths (MoED), a novel architectural framework that synergistically unifies dynamic width and depth scaling. MoED employs a hierarchical routing mechanism, where a meta-controller at each layer makes a joint decision on both expert selection and a token's subsequent computational path?whether to exit, proceed, or skip ahead. This approach creates a unique, input-adaptive sub-network for every token, optimizing the allocation of compute. The proposed architecture presents a fundamental shift towards more efficient and scalable LLMs, theoretically enabling superior performance and reduced latency while managing the activation memory bottlenecks that plague traditional trillion-parameter models.},
      keywords = {Large Language Models, Mixture of Experts, Adaptive Computation, Dynamic Networks, Hierarchical Routing, Efficient Inference},
      month = {August},
  }