mlsanityData to 14 Oct 2026

Task-Aware QUBO Allocation for Mixed-Precision Quantization

arXiv 2609.22238 ↗First posted 4 Sep 2026cs.LG cs.CVcited by 0

Not on the map. Why?

Abstract (arXiv)

Mixed-precision quantization requires discrete allocation of weight and activation bit-widths, followed by recovery of the selected network. We develop a task-aware quadratic unconstrained binary optimization (QUBO) surrogate with separate weight and activation profiles, a bit-operation (BOP) cost, and selected structural priors. QUBO provides a network-wide allocation that can be refined through direct validation-based PROTES search. On a compact NAFBlock-based denoiser, the refined route achieves 37.192 dB after LSQ+ at 4.035% routed-layer BOPs, versus 37.092 dB at 4.101% for a HAWQ-style baseline. The repeated-search primary experiment shows that LSQ+ largely closes the quality gap between QUBO allocation and expensive direct refinement. An additional restoration architecture retains a larger recovered gain, indicating that refinement's value depends on architecture and recovery. We evaluate quality, achieved cost, routing stability and optimization expense together. Deployment measurements characterize a fake-quantized floating-point implementation; BOP reductions describe analytical allocation savings.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 14 of its 22 references that are in the map. First posted on arXiv 4 Sep 2026.

Builds on

14 of 22 references.

References · 14

  1. 2018MobileNetV2: Inverted Residuals and Linear Bottleneckscited by 4,795
  2. 2022Simple Baselines for Image Restorationcited by 708
  3. 2019Learned Step Size Quantizationcited by 628
  4. 2020Up or Down? Adaptive Rounding for Post-Training Quantizationcited by 621
  5. 2021BRECQ: Pushing the Limit of Post-Training Quantization by Block Reconstructioncited by 479
  6. 2018PACT: Parameterized Clipping Activation for Quantized Neural Networkscited by 467
  7. 2018HAQ: Hardware-Aware Automated Quantization with Mixed Precisioncited by 443
  8. 2019HAWQ: Hessian AWare Quantization of Neural Networks with Mixed-Precisioncited by 402
  9. 2019HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networkscited by 261
  10. 2020LSQ+: Improving low-bit quantization through learnable offsets and better initializationcited by 189
  11. 2020Rethinking Differentiable Search for Mixed-Precision Neural Networkscited by 82
  12. 2020Differentiable Joint Pruning and Quantization for Hardware Efficiencycited by 46
2 more
  1. 2025Is Quantum Optimization Ready? An Effort Towards Neural Network Compression using Adiabatic Quantum Computingcited by 2
  2. 2025Optimization of the quantization of dense neural networks from an exact QUBO formulationcited by 2

Built on by

Not cited yet.

Similar, not cited

Close in wording, no citation either way

  1. 2021Bit-Mixer: Mixed-precision networks with runtime bit-width selection
  2. 2026Beyond Outliers: A Data-Free Layer-wise Mixed-Precision Quantization Approach Driven by Numerical and Structural Dual-Sensitivity
  3. 2023Free Bits: Latency Optimization of Mixed-Precision Quantized Neural Networks on the Edge
  4. 2023On-Chip Hardware-Aware Quantization for Mixed Precision Neural Networks
  5. 2026StableQAT: Stable Quantization-Aware Training at Ultra-Low Bitwidths
  6. 2023SONIQ: System-Optimized Noise-Injected Ultra-Low-Precision Quantization with Full-Precision Parity
  7. 2026Beyond Scalar Sensitivity: Activation-Aware Mixed-Precision LLM Quantization with Cross-Layer Refinement
  8. 2026Q-PACE: Dynamic Precision Allocation for Quantization-Aware Training