OTCE: Hybrid SSM and Attention with Cross Domain Mixture of Experts to construct Observer-Thinker-Conceiver-Expresser

TL;DR

This paper introduces OTCE, a hybrid architecture combining selective state space models and quadratic self-attention with cross-domain experts, achieving competitive performance in language modeling tasks.

Contribution

It proposes a novel biomimetic architecture that integrates state space models and quadratic attention through hybrid experts with cross-sharing domains.

Findings

OTCE competes with medium-scale open-source language models.

The hybrid architecture effectively combines advantages of state space and attention mechanisms.

Position information injection enhances long-term dependency handling.

Abstract

Recent research has shown that combining Mamba with Transformer architecture, which has selective state space and quadratic self-attention mechanism, outperforms using Mamba or Transformer architecture alone in language modeling tasks. The quadratic self-attention mechanism effectively alleviates the shortcomings of selective state space in handling long-term dependencies of any element in the sequence. We propose a position information injection method that connects the selective state space model with the quadratic attention, and integrates these two architectures with hybrid experts with cross-sharing domains, so that we can enjoy the advantages of both. We design a new architecture with a more biomimetic idea: Observer-Thinker-Conceiver-Expresser (OTCE), which can compete with well-known medium-scale open-source language models on a small scale in language modeling tasks.