Pith. sign in

REVIEW 8 cited by

Mixture of A Million Experts

Not yet reviewed by Pith; the record is open.

This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.

SPECIMEN: schema-true, not a live event

T0 review · schema-true

One-sentence machine reading of the paper's core claim.

pith:XXXXXXXX · record.json · timestamp

arxiv 2407.04153 v1 pith:RPK6NES4 submitted 2024-07-04 cs.LG cs.AI

classification cs.LGcs.AI
keywords computationalexpertspeerarchitecturesefficientlayerlayersmillion
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

The feedforward (FFW) layers in standard transformer architectures incur a linear increase in computational costs and activation memory as the hidden layer width grows. Sparse mixture-of-experts (MoE) architectures have emerged as a viable approach to address this issue by decoupling model size from computational cost. The recent discovery of the fine-grained MoE scaling law shows that higher granularity leads to better performance. However, existing MoE models are limited to a small number of experts due to computational and optimization challenges. This paper introduces PEER (parameter efficient expert retrieval), a novel layer design that utilizes the product key technique for sparse retrieval from a vast pool of tiny experts (over a million). Experiments on language modeling tasks demonstrate that PEER layers outperform dense FFWs and coarse-grained MoEs in terms of performance-compute trade-off. By enabling efficient utilization of a massive number of experts, PEER unlocks the potential for further scaling of transformer models while maintaining computational efficiency.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 8 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. Training, Reading, and Editing Legible Transformers

    cs.LG 2026-07 conditional novelty 6.5 of 10

    A variance-floor objective plus learned operator gates produce an end-to-end legible transformer whose crisp units are 50–184× more local to edit and can be reshaped from fan-out to fan-in circuits without quality loss.

  2. OmniMoE: An Efficient MoE by Orchestrating Atomic Experts at Scale

    cs.CL 2026-02 conditional novelty 6.0 of 10

    OmniMoE's vector-level atomic experts + Cartesian-product router + expert-centric scheduling gives 50.9% zero-shot accuracy and a 10.9x inference speedup over PEER at matched 1.7B active parameters.

  3. The SpiNNaker2 chip: a many-core platform for flexible and scalable brain-inspired computing

    cs.ET 2026-07 accept novelty 5.5 of 10

    SpiNNaker2 delivers a measured many-core platform combining ARM cores, ML accelerators, and event routing that runs SNNs, DNNs, and hybrid event-based models on one scalable chip.

  4. More Than Memory: Task-Conditioned Signed FFN Writes in Long-Context Retrieval

    cs.LG 2026-06 conditional novelty 5.0 of 10

    Scaling a model's own per-layer feed-forward writes shows signed, task-dependent retrieval roles, and the sign of the write–gradient dot product identifies suppressors from amplifiers.

  5. MoLEx: Mixture of LoRA Experts in Speech Self-Supervised Models for Audio Deepfake Detection

    cs.SD 2025-09 conditional novelty 5.0 of 10

    MoLEx combines LoRA adapters with a top-K expert router inside a frozen WavLM model, achieving 5.56% EER on ASVSpoof 5 without augmentation.

  6. PC-MoE: Memory-Efficient and Privacy-Preserving Collaborative Training for Mixture-of-Experts LLMs

    cs.LG 2025-06 conditional novelty 5.0 of 10

    PC-MoE shards the expert layers of an MoE LLM across parties and routes only sparse top-k activations between them, achieving near-centralized accuracy with about 70% memory savings and resistance to one partial-gradi...

  7. UltraMemV2: Memory Networks Scaling to 120B Parameters with Superior Long-Context Learning

    cs.LG 2025-08 conditional novelty 4.0 of 10

    A redesigned memory-layer architecture with five engineering improvements reaches performance parity with 8-expert MoE at similar compute, with lower memory access and stronger long-context memorization.

  8. Scaling and Enhancing LLM-based AVSR: A Sparse Mixture of Projectors Approach

    eess.AS 2025-05 conditional novelty 4.0 of 10

    Llama-SMoP-DEDR, a sparse mixture of projectors with modality-specific experts and routers, lowers word error rate for LLM-based AVSR on LRS3, mainly with smaller LLMs.

Pith tools