arXiv:2606. 04980v1 Announce Type: new Abstract: Mixture-of-Experts (MoE) architectures scale model capacity through sparse expert activation, but their deployment remains memory-bound because all expert weights must reside in memory.
arXiv:2606. 00079v1 Announce Type: cross Abstract: Mixture-of-Experts (MoE) large language models reduce per-token computation through sparse expert activation, but their deployment remains memory-intensive because all expert weights must be kept resident in memory.
The paper introduces Colla-Q, a Mixture-of-Experts (MoE) quantization technique that uses activation entropy to allocate bit-widths across experts. By balancing performance among experts, Colla-Q improves overall MoE accuracy and reduces reliance on calibration datasets. The method aims to maintain robustness and stability in quantized MoE models.
The paper introduces a curiosity‑driven quantized Mixture‑of‑Experts framework that routes inputs based on Bayesian epistemic uncertainty across heterogeneous experts (BitNet ternary, 1‑16 bit BitLinear, post‑training quantization). On audio classification benchmarks, 4‑bit quantization preserves 99.9 % of full‑precision F1 while achieving 4× compression and 31 % energy savings, and curiosity‑driven routing further improves accuracy and reduces cross‑fold variance by up to 85 %. The routing is self‑organizing, allocating the most uncertain samples to the high‑precision expert, and the method demonstrates statistical parity with full precision across datasets.
Dynamic Expert Quantization (DynaExq) is a runtime-aware mixed-precision serving system designed for single‑GPU Mixture‑of‑Experts (MoE) inference under a hard high‑bandwidth memory (HBM) envelope. It treats the problem as an online, budget‑constrained precision allocation task, keeping the most frequently used experts at higher precision while relegating the rest to low‑precision fallbacks. By estimating expert hotness from router traces and asynchronously promoting or demoting experts, DynaExq maintains a fully materialized expert set during the forward pass, improving accuracy and throughput compared to static post‑training quantization and offloading/prefetch baselines. whyItMatters":"DynaExq enables efficient deployment of large MoE models on memory‑limited GPUs by dynamically allocating precision based on runtime expert usage, thereby reducing memory footprint and latency while boosting accuracy and throughput."
arXiv:2605. 26660v2 Announce Type: replace Abstract: Quantization is an effective approach to reduce the memory footprint and inference cost of large language models (LLMs), yet maintaining performance in the ultra-low-bit regime remains challenging.
arXiv:2602. 06154v2 Announce Type: replace Abstract: Mixture-of-Experts (MoE) models scale large language models efficiently by sparsely activating experts, but once an expert is selected, it is executed fully.
arXiv:2607. 23047v1 Announce Type: cross Abstract: Mixed-precision quantization improves the accuracy of post-training quantization by allocating higher bitwidths to sensitive layers, but existing methods solve the allocation for a single fixed memory budget.
GAMMA is a post‑training framework that learns module‑wise precision preferences for mixed‑precision quantization of large language models. It optimizes a teacher‑forced hidden‑state reconstruction objective under an augmented Lagrangian constraint and then projects the learned preferences into exact budget‑feasible discrete assignments via integer programming. Because the learned preferences encode a stable sensitivity ranking, a single training run can be reused for any deployment budget, reducing per‑budget adaptation from hours to minutes and outperforming fixed‑precision baselines and search‑based methods on Llama and Qwen models.
arXiv:2609.36222v1 Announce Type: new Abstract: Large language models are increasingly expensive to serve. In large-scale serving systems, autoregressive decoding is often bottlenecked by transferrin...
arXiv:2606. 04115v1 Announce Type: cross Abstract: Quantizing large language models (LLMs) to low-precision floating-point representations is central to efficient deployment, yet applying a single bit-width uniformly across all layers is sub-optimal in terms of both performance and accuracy.
arXiv:2606. 04238v1 Announce Type: cross Abstract: Aggressive weight quantization to 2-bit precision offers substantial throughput and memory gains for large language model (LLM) inference, but typically incurs severe accuracy degradation.
arXiv:2608. 05499v1 Announce Type: cross Abstract: Modern deep neural networks achieve strong performance, but their scale makes them costly and slow, especially on resource-constrained edge devices.