EdgeXpert: An Edge Device for Memory-Efficient LLM Inference with Mixture-of-Experts and Speculative Decoding

arXiv:2608.05303 · cs.AR, cs.CL, cs.LG · Submitted 2026-08-05 · Read on arXiv

Sangwoo Ha, Hyunwoo Seo, Yurim Jo, Youngjin Moon, Hoi-Jun Yoo

cs.AR, cs.CL, cs.LG

Submitted: 2026-08-05

Comments: Accepted at the 59th IEEE/ACM International Symposium on Microarchitecture (MICRO 2026)

Code: https://github.com/vllm-project/llm-compressor

License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/

Terminology

• Speculative Decoding — 77× in this paper · explained in 2 episodes
• Pruning — 64× in this paper · explained in 4 episodes
• Accuracy — 23× in this paper · explained in 1 episode
• Decode — 22× in this paper · explained in 2 episodes
• Verification — 21× in this paper · explained in 2 episodes
• Edge Devices — 18× in this paper · explained in 2 episodes
• Calibration — 15× in this paper · explained in 9 episodes
• Transformer — 14× in this paper · explained in 2 episodes
• Quantization — 10× in this paper · explained in 15 episodes
• Benchmark — 9× in this paper · explained in 2 episodes
• Sparsity — 9× in this paper · explained in 5 episodes
• Transformers — 8× in this paper · explained in 2 episodes
• Core — 5× in this paper · explained in 2 episodes
• Edge Deployment — 4× in this paper · explained in 2 episodes
• Allocator — 3× in this paper · explained in 2 episodes
• Generator — 3× in this paper · explained in 2 episodes
• Hyperparameters — 3× in this paper · explained in 2 episodes
• KV cache — 3× in this paper · explained in 4 episodes
• Mixture-of-Experts (MoE) — 3× in this paper · explained in 11 episodes
• Annual International Symposium on Computer Architecture — 2× in this paper
• Decoder — 2× in this paper · explained in 2 episodes
• Energy Efficiency — 2× in this paper · explained in 4 episodes
• IEEE International Solid-State Circuits Conference — 2× in this paper
• Language Models (LLMs) — 2× in this paper · explained in 2 episodes
• Large Language Models (LLMs) — 2× in this paper · explained in 9 episodes
• Mixture of Experts — 2× in this paper · explained in 2 episodes
• Papers — 2× in this paper · explained in 3 episodes
• Thinking mode — 2× in this paper · explained in 2 episodes
• Trace — 2× in this paper · explained in 2 episodes
• Consistency — 1× in this paper · explained in 3 episodes
• Continuous Batching — 1× in this paper · explained in 2 episodes
• Fine-Tuning — 1× in this paper · explained in 5 episodes
• International Symposium on High Performance Computer Architecture — 1× in this paper
• LLM-as-a-Judge — 1× in this paper · explained in 4 episodes
• Mixture-of-Experts (MoE) Models — 1× in this paper · explained in 3 episodes
• Model Compression — 1× in this paper · explained in 3 episodes
• Perplexity — 1× in this paper · explained in 3 episodes
• RAG — 1× in this paper · explained in 3 episodes
• Redundancy — 1× in this paper · explained in 2 episodes
• Reinforcement Learning — 1× in this paper · explained in 4 episodes
• Sparse Attention — 1× in this paper · explained in 5 episodes
• Transformer Architecture — 1× in this paper · explained in 4 episodes
• Verifier — 1× in this paper · explained in 2 episodes
• Vision-language models — 1× in this paper · explained in 2 episodes

Sources

• MoE-Pruner: Pruning Mixture-of-Experts Large Language Model using the Hints from Its Router
• GPT-4 Technical Report
• The Llama 3 Herd of Models
• Gemini 1.5: Unlocking multimodal understanding across millions of tokens of context
• Squeezed Attention: Accelerating Long Context Length LLM Inference
• Qwen3 Technical Report
• DeepSeekMoE: Towards Ultimate Expert Specialization in Mixture-of-Experts Language Models
• Mixtral of Experts
• Medusa: Simple LLM Inference Acceleration Framework with Multiple Decoding Heads
• Speculative Decoding: Exploiting Speculative Execution for Accelerating Seq2seq Generation
• Accelerating Large Language Model Decoding with Speculative Sampling
• LayerSkip: Enabling Early Exit Inference and Self-Speculative Decoding
• EAGLE-3: Scaling up Inference Acceleration of Large Language Models via Training-Time Test
• ML-SpecQD: Multi-Level Speculative Decoding with Quantized Drafts
• SpecMemo: Speculative Decoding is in Your Pocket
• DeepSeek-R1: Incentivizing Reasoning Capability in LLMs via Reinforcement Learning
• EAGLE: Speculative Sampling Requires Rethinking Feature Uncertainty
• MoESD: Unveil Speculative Decoding's Potential for Accelerating Sparse MoE
• DeepSeek-V2: A Strong, Economical, and Efficient Mixture-of-Experts Language Model
• OLMoE: Open Mixture-of-Experts Language Models

Related papers

• WitCert: Sound Runtime Risk Observability and Gating for KV-Cache Quantization
• Golden Ruler: A Numeric Format Catalog with Bit-Exact Conformance Vectors for FP8, BF16, MXFP4, and Microscaling Formats
• RFChipAgent: Multi-Agentic AI Flow for Analog/RF Chip Design
• PoisonCap: Efficient Hierarchical Temporal Safety for CHERI
• Provisioning to Runtime Optimization of a 100 MW-Scale AI Cluster
• Bit-Accurate Modeling of GPU Matrix Multiply-Accumulate Units: Demystifying Numerical Discrepancy and Accuracy