# ExpertFlow: Adaptive Expert Scheduling and Memory Coordination for Efficient MoE Inference

arXiv: https://arxiv.org/abs/2510.26730 · HTML: https://mlsanity.com/paper/2510.26730/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- Placed in Efficient MoE training and inference systems, one of 12 directions in Mixture-of-experts.
- Builds on 16 of its 42 references that are in the map.
- First posted on arXiv 30 Oct 2025.

## Placed in

- Mixture-of-experts / [Efficient MoE training and inference systems](https://mlsanity.com/moe/efficient-moe-training-and-inference-systems/)

## Builds on (most cited first, 16 of 16)

- [Qwen2.5 Technical Report](https://mlsanity.com/paper/2412.15115/) (2024)
- [Qwen2 Technical Report](https://mlsanity.com/paper/2407.10671/) (2024)
- [DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model](https://mlsanity.com/paper/2405.04434/) (2024)
- [DeepSeek LLM: Scaling Open-Source Language Models with Longtermism](https://mlsanity.com/paper/2401.02954/) (2024)
- [Adaptive Computation Time for Recurrent Neural Networks](https://mlsanity.com/paper/1603.08983/) (2016)
- [Time-MoE: Billion-Scale Time Series Foundation Models with Mixture of Experts](https://mlsanity.com/paper/2409.16040/) (2024)
- [Fast Inference of Mixture-of-Experts Language Models with Offloading](https://mlsanity.com/paper/2312.17238/) (2023)
- [MoE-Lightning: High-Throughput MoE Inference on Memory-constrained GPUs](https://mlsanity.com/paper/2411.11217/) (2024)
- [HOBBIT: A Mixed Precision Expert Offloading System for Fast MoE Inference](https://mlsanity.com/paper/2411.01433/) (2024)
- [AdaMoE: Token-Adaptive Routing with Null Experts for Mixture-of-Experts Language Models](https://mlsanity.com/paper/2406.13233/) (2024)
- [ProMoE: Fast MoE-based LLM Serving using Proactive Caching](https://mlsanity.com/paper/2410.22134/) (2024)
- [ExpertFlow: Efficient Mixture-of-Experts Inference via Predictive Expert Caching and Token Scheduling](https://mlsanity.com/paper/2410.17954/) (2024)
- [Leveraging Mixture of Experts for Improved Speech Deepfake Detection](https://mlsanity.com/paper/2409.16077/) (2024)
- [A Gated Residual Kolmogorov-Arnold Networks for Mixtures of Experts](https://mlsanity.com/paper/2409.15161/) (2024)
- [Mixture of Diverse Size Experts](https://mlsanity.com/paper/2409.12210/) (2024)
- [Mixture of Experts for Node Classification](https://mlsanity.com/paper/2412.00418/) (2024)

## Built on by (most cited first, 5 of 5)

- [HUD: Hierarchical Uncertainty-Aware Disambiguation Network for Composed Video Retrieval](https://mlsanity.com/paper/2512.02792/) (2025)
- [SCRIBE: Structured Mid-Level Supervision for Tool-Using Language Models](https://mlsanity.com/paper/2601.03555/) (2026)
- [Dynamic Expert Quantization for Scalable Mixture-of-Experts Inference](https://mlsanity.com/paper/2511.15015/) (2025)
- [VisMMOE: Exploiting Visual-Expert Affinity for Efficient Visual-Language MoE Offloading](https://mlsanity.com/paper/2605.05899/) (2026)
- [Beyond Independent Optimization: Compression, MoE Routing, and Quantization Interactions in Multimodal Edge Intelligence](https://mlsanity.com/paper/2607.20981/) (2026)

## Abstract (arXiv)

The expansion of large language models is increasingly limited by the constrained memory capacity of modern GPUs. To mitigate this, Mixture-of-Experts (MoE) architectures activate only a small portion of parameters during inference, significantly lowering both memory demand and computational overhead. However, conventional MoE inference approaches, which select active experts independently at each layer, often introduce considerable latency because of frequent parameter transfers between host and GPU memory. In addition, current cross-layer prediction strategies, which are typically based on fixed steps, lack adaptability across different hardware platforms and workloads, thereby reducing their robustness and effectiveness. To address these challenges, we present ExpertFlow, a runtime system for MoE inference that combines adaptive expert prefetching and cache-aware routing. ExpertFlow continuously adjusts its prediction horizon for expert activation by leveraging runtime statistics such as transfer bandwidth, parameter dimensionality, and model feedback signals. Furthermore, it incorporates a hybrid cross-layer prediction scheme that fuses pregating information with intermediate computational states to anticipate future expert needs. By adaptively refining prefetching decisions and aligning them with actual usage behavior, ExpertFlow effectively decreases cache misses and removes latency caused by expert swap-ins. Our evaluation demonstrates that ExpertFlow reduces model stall time to less than 0.1% of the baseline, highlighting its capability to optimize MoE inference under stringent memory constraints.