Tootfinder

ButterflyQuant: Ultra-low-bit LLM Quantization through Learnable Orthogonal Butterfly Transforms
Bingxin Xu, Zhen Dong, Oussama Elachqar, Yuzhang Shang
https://arxiv.org/abs/2509.09679

ButterflyQuant: Ultra-low-bit LLM Quantization through Learnable Orthogonal Butterfly Transforms
Large language models require massive memory footprints, severely limiting deployment on consumer hardware. Quantization reduces memory through lower numerical precision, but extreme 2-bit quantization suffers from catastrophic performance loss due to outliers in activations. Rotation-based methods such as QuIP and QuaRot apply orthogonal transforms to eliminate outliers before quantization, using computational invariance: $\mathbf{y} = \mathbf{Wx} = (\mathbf{WQ}^T)(\mathbf{Qx})$ for orthogonal…

Spatial-Temporal Multi-Scale Quantization for Flexible Motion Generation
Zan Wang, Jingze Zhang, Yixin Chen, Baoxiong Jia, Wei Liang, Siyuan Huang
https://arxiv.org/abs/2508.08991

Spatial-Temporal Multi-Scale Quantization for Flexible Motion Generation
Despite significant advancements in human motion generation, current motion representations, typically formulated as discrete frame sequences, still face two critical limitations: (i) they fail to capture motion from a multi-scale perspective, limiting the capability in complex patterns modeling; (ii) they lack compositional flexibility, which is crucial for model's generalization in diverse generation tasks. To address these challenges, we introduce MSQ, a novel quantization method that compre…

Finite Scalar Quantization Enables Redundant and Transmission-Robust Neural Audio Compression at Low Bit-rates
Harry Julia, Rachel Beeson, Lohith Konathala, Johanna Ulin, Jiameng Gao
https://arxiv.org/abs/2509.09550

Finite Scalar Quantization Enables Redundant and Transmission-Robust Neural Audio Compression at Low Bit-rates
Neural Audio Codecs (NACs) have become increasingly adopted in speech processing tasks due to their excellent rate-distortion performance and compatibility with Large Language Models (LLMs) as discrete feature representations for audio generation. While most existing codecs rely on Residual Vector Quantization (RVQ), Finite Scalar Quantization (FSQ) has recently emerged as a compelling alternative that simplifies training and natively supports single codebooks. We introduce NeuCodec, an FSQ-bas…

HERO: Hardware-Efficient RL-based Optimization Framework for NeRF Quantization
Yipu Zhang, Chaofang Ma, Jinming Ge, Lin Jiang, Jiang Xu, Wei Zhang
https://arxiv.org/abs/2510.09010

HERO: Hardware-Efficient RL-based Optimization Framework for NeRF Quantization
Neural Radiance Field (NeRF) has emerged as a promising 3D reconstruction method, delivering high-quality results for AR/VR applications. While quantization methods and hardware accelerators have been proposed to enhance NeRF's computational efficiency, existing approaches face crucial limitations. Current quantization methods operate without considering hardware architecture, resulting in sub-optimal solutions within the vast design space encompassing accuracy, latency, and model size. Additio…

Quantization of charged fields in the presence of intense electromagnetic fields
\'Alvaro \'Alvarez-Dom\'inguez
https://arxiv.org/abs/2510.09447 https://

Quantization of charged fields in the presence of intense electromagnetic fields
This thesis applies techniques from quantum field theory in curved spacetimes to study particle creation in external fields, focusing on the Schwinger effect (i.e., the production of particle-antiparticle pairs by intense electric fields). Although experimental verification remains out of reach, theoretical analysis advances our understanding of this phenomenon and its broader implications. The work develops the theoretical framework for quantizing charged fields in nontrivial backgrounds, ad…

Profiling Large Language Model Inference on Apple Silicon: A Quantization Perspective
Afsara Benazir, Felix Xiaozhu Lin
https://arxiv.org/abs/2508.08531 https://

Profiling Large Language Model Inference on Apple Silicon: A Quantization Perspective
A systematic understanding of Apple Silicon is lacking in the current landscape of hardware efficiency; research focus is largely centered on accelerating GPUs for large-scale training or inference on CUDA devices. This paper investigates Apple Silicon's unique memory architecture that offers a unified memory integrating CPU and GPU memory and its implications for on-device LLM inference. We decipher myths about whether Apple Silicon is efficient for on-device inference compared to competitor…

Quantization of the electromagnetic fields from single atomic or molecular radiators
Valerica Raicu
https://arxiv.org/abs/2509.07359 https://arxiv.org/pdf/…

Quantization of the electromagnetic fields from single atomic or molecular radiators
A framework is introduced for expressing electromagnetic (EM) potentials and fields of single atomic or molecular emitters modeled as oscillating dipoles, which follows a recently proposed method for solving inhomogeneous wave equations for arbitrary, time-dependent distributions of charge. This framework is first used to evaluate the physical implications of simplifying assumptions made in the standard approach to quantization of the EM fields and the impact of such assumptions on the results …

Quantization of bounded symplectic domains associated with compact Lie groups
Alexey A. Sharapov
https://arxiv.org/abs/2509.05931 https://arxiv.org/pdf/250…

Quantization of bounded symplectic domains associated with compact Lie groups
We present a systematic quantization scheme for bounded symplectic domains of the form $D \times G \subset T^\ast G$, where $D \subset \mathfrak{g}^\ast$ is a bounded region defined by algebraic inequalities and $G$ is a compact Lie group with Lie algebra $\mathfrak{g}$. The finiteness of the symplectic volume implies that quantization yields a finite-dimensional Hilbert space, with observables represented by Hermitian matrices, for which we provide an explicit realization. Boundary effects nec…

Training Dynamics Impact Post-Training Quantization Robustness
Albert Catalan-Tatjer, Niccol\`o Ajroldi, Jonas Geiping
https://arxiv.org/abs/2510.06213 https://

Training Dynamics Impact Post-Training Quantization Robustness
While post-training quantization is widely adopted for efficient deployment of large language models, the mechanisms underlying quantization robustness remain unclear. We conduct a comprehensive analysis of quantization degradation across open-source language model training trajectories up to 32B parameters and 15T training tokens to accurately assess the relationship between training dynamics and quantization performance. Our key finding is that quantization errors in large-scale training runs…

Hierarchical Spatial Algorithms for High-Resolution Image Quantization and Feature Extraction
Noor Islam S. Mohammad
https://arxiv.org/abs/2510.08449 https://

Hierarchical Spatial Algorithms for High-Resolution Image Quantization and Feature Extraction
This study introduces a modular framework for spatial image processing, integrating grayscale quantization, color and brightness enhancement, image sharpening, bidirectional transformation pipelines, and geometric feature extraction. A stepwise intensity transformation quantizes grayscale images into eight discrete levels, producing a posterization effect that simplifies representation while preserving structural detail. Color enhancement is achieved via histogram equalization in both RGB and Y…

Recursive algorithm for constructing antisymmetric fermionic states in first quantization mapping
E. Rule, I. A. Chernyshev, I. Stetcu, J. Carlson, R. Weiss
https://arxiv.org/abs/2509.07279

Recursive algorithm for constructing antisymmetric fermionic states in first quantization mapping
We devise a deterministic quantum algorithm to produce antisymmetric states of single-particle orbitals in the first quantization mapping. Unlike sorting-based antisymmetrization algorithms, which require ordered input states and high Clifford-gate overhead, our approach initializes the state of each particle independently. For a system of $N$ particles and $N_s$ single-particle states, our algorithm prepares antisymmetrized states of non-trivial localized (e.g., Hartree-Fock) orbitals using $O…

Sharpness-Aware Data Generation for Zero-shot Quantization
Dung Hoang-Anh, Cuong Pham Trung Le, Jianfei Cai, Thanh-Toan Do
https://arxiv.org/abs/2510.07018 https://

Sharpness-Aware Data Generation for Zero-shot Quantization
Zero-shot quantization aims to learn a quantized model from a pre-trained full-precision model with no access to original real training data. The common idea in zero-shot quantization approaches is to generate synthetic data for quantizing the full-precision model. While it is well-known that deep neural networks with low sharpness have better generalization ability, none of the previous zero-shot quantization works considers the sharpness of the quantized model as a criterion for generating tr…

Optimizing Fronthaul Quantization for Flexible User Load in Cell-Free Massive MIMO
Fabian G\"ottsch, Max Franke, Arash Pourdamghani, Giuseppe Caire, Stefan Schmid
https://arxiv.org/abs/2510.06734 …

Optimizing Fronthaul Quantization for Flexible User Load in Cell-Free Massive MIMO
We investigate the physical layer (PHY) spectral efficiency and fronthaul network load of a scalable user-centric cell-free massive MIMO system. Each user-centric cluster processor responsible for cluster-level signal processing is located at one of multiple decentralized units (DUs). Thus, the radio units in the cluster must exchange data with the corresponding DU over the fronthaul. Because the fronthaul links have limited capacity, this data must be quantized before it is sent over the front…

Rate-Adaptive Semantic Communication via Multi-Stage Vector Quantization
Jinsung Park, Junyong Shin, Yongjeong Oh, Jihun Park, Yo-Seb Jeon
https://arxiv.org/abs/2510.02646 https…

Rate-Adaptive Semantic Communication via Multi-Stage Vector Quantization
This paper proposes a novel framework for rate-adaptive semantic communication based on multi-stage vector quantization (VQ), termed \textit{MSVQ-SC}. Unlike conventional single-stage VQ approaches, which require exponentially larger codebooks to achieve higher fidelity, the proposed framework decomposes the quantization process into multiple stages and dynamically activates both stages and individual VQ modules. This design enables fine-grained rate adaptation under varying bit constraints whi…

Supersymmetric lattice theories on curved space
David Berenstein, Simon Catterall
https://arxiv.org/abs/2509.08885 https://arxiv.org/pdf/2509.08885

Supersymmetric lattice theories on curved space
We show how to construct Hamiltonian lattice theories with one exact supersymmetry on arbitrary triangulations of curved space in any number of dimensions. Both bosons and fermions satisfy discrete Kähler-Dirac equations. The quantization of the fermions proceeds by imposing conventional anti-commutation relations while the bosons require a modification of the usual canonical commutator. On regular lattices we construct parity, time reversal and translation-by-one (shift) symmetries. We argue …

The Uneven Impact of Post-Training Quantization in Machine Translation
Benjamin Marie, Atsushi Fujita
https://arxiv.org/abs/2508.20893 https://arxiv.org/pd…

The Uneven Impact of Post-Training Quantization in Machine Translation
Quantization is essential for deploying large language models (LLMs) on resource-constrained hardware, but its implications for multilingual tasks remain underexplored. We conduct the first large-scale evaluation of post-training quantization (PTQ) on machine translation across 55 languages using five LLMs ranging from 1.7B to 70B parameters. Our analysis reveals that while 4-bit quantization often preserves translation quality for high-resource languages and large models, significant degradati…

The Schr\"odinger equation for a spherically symmetric system, its structure and solutions
R. I. Ayala O\~na, T. P. Shestakova
https://arxiv.org/abs/2509.08523 https://

The Schrödinger equation for a spherically symmetric system, its structure and solutions
The Wheeler - DeWitt geometrodynamics, as the first attempt to develop a quantum theory of gravity, faces certain challenges, including the problem of time and the interpretation of the wave function. In this paper, we present the extended phase space approach to quantization of gravity as an alternative approach to the Wheeler - DeWitt quantum geometrodynamics. For a spacetime with a nontrivial topology, the Wheeler - DeWitt equation loses its sense, but we can derive the Schrödinger equation…

CSI Compression Beyond Latents: End-to-End Hybrid Attention-CNN Networks with Entropy Regularization
Maryam Ansarifard, Mostafa Rahmani, Mohit K. Sharma, Kishor C. Joshi, George Exarchakos, Alister Burr
https://arxiv.org/abs/2509.08776

CSI Compression Beyond Latents: End-to-End Hybrid Attention-CNN Networks with Entropy Regularization
Massive MIMO systems rely on accurate Channel State Information (CSI) feedback to enable high-gain beam- forming. However, the feedback overhead scales linearly with the number of antennas, presenting a major bottleneck. While recent deep learning methods have improved CSI compres- sion, most overlook the impact of quantization and entropy coding, limiting their practical deployability. In this work, we propose an end-to-end CSI compression framework that inte- grates a Spatial Correlation-Guid…

LIV-Decoherence on Gravitational Cat States
Iarley P. Lobo, Kelvin Sampaio, Gislaine Var\~ao, Moises Rojas, Valdir B. Bezerra
https://arxiv.org/abs/2510.08828 https://

LIV-Decoherence on Gravitational Cat States
Inspired by approaches based on the stochastic generalized uncertainty principle, we propose a Lindblad equation derived from the quantization of a stochastic modified dispersion relation in a Lorentz Invariance Violation (LIV) scenario. This framework enables us to investigate decoherence effects in a system of particles exhibiting gravitationally induced entanglement. We analyze the impact of LIV on entanglement (quantified by concurrence) considering systematic and stochastic effects.

Quantization of spin circular photogalvanic effect in altermagnetic Weyl semimetals
Hiroki Yoshida, Jan Priessnitz, Libor \v{S}mejkal, Shuichi Murakami
https://arxiv.org/abs/2509.05620

Quantization of spin circular photogalvanic effect in altermagnetic Weyl semimetals
We theoretically predict a spin-current analog of the quantized circular photogalvanic effect in Weyl semimetals. This phenomenon is forbidden in antiferromagnets by symmetry but uniquely allowed in altermagnets, highlighting a novel and intrinsic characteristic of altermagnetism. To systematically explore second-order spin current responses, we classify all symmetry-allowed responses based on spin point groups. Furthermore, we provide a comprehensive classification of altermagnetic Weyl semime…

Duality between polyhedral approximation of value functions and optimal quantization of measures
Abdellah Bulaich Mehamdi, Wim van Ackooij, Luce Brotcorne, St\'ephane Gaubert, Quentin Jacquet
https://arxiv.org/abs/2509.04101

Duality between polyhedral approximation of value functions and optimal quantization of measures
Approximating a convex function by a polyhedral function that has a limited number of facets is a fundamental problem with applications in various fields, from mitigating the curse of dimensionality in optimal control to bi-level optimization. We establish a connection between this problem and the optimal quantization of a positive measure. Building on recent stability results in optimal transport, by Delalande and Mérigot, we deduce that the polyhedral approximation of a convex function is eq…

Huawei's Zurich Lab unveils SINQ, an open-source quantization method that it claims can reduce LLM memory use by 60-70% without significant quality loss (Carl Franzen/VentureBeat)
https://venturebeat.com/ai/huaweis-new-open-source-techni…

Integrating Pruning with Quantization for Efficient Deep Neural Networks Compression
Sara Makenali, Babak Rokh, Ali Azarpeyvand
https://arxiv.org/abs/2509.04244 https://

Integrating Pruning with Quantization for Efficient Deep Neural Networks Compression
Deep Neural Networks (DNNs) have achieved significant advances in a wide range of applications. However, their deployment on resource-constrained devices remains a challenge due to the large number of layers and parameters, which result in considerable computational and memory demands. To address this issue, pruning and quantization are two widely used compression techniques, commonly applied individually in most studies to reduce model size and enhance processing speed. Nevertheless, combining…

Robust Residual Finite Scalar Quantization for Neural Compression
Xiaoxu Zhu
https://arxiv.org/abs/2508.15860 https://arxiv.org/pdf/2508.15860

Robust Residual Finite Scalar Quantization for Neural Compression
Finite Scalar Quantization (FSQ) has emerged as a promising alternative to Vector Quantization (VQ) in neural compression, offering simplified training and improved stability. However, naive application of FSQ in residual quantization frameworks suffers from the \textbf{residual magnitude decay problem}, where subsequent FSQ layers receive progressively weaker signals, severely limiting their effectiveness. We propose \textbf{Robust Residual Finite Scalar Quantization (RFSQ)}, a general framewo…

FineServe: Precision-Aware KV Slab and Two-Level Scheduling for Heterogeneous Precision LLM Serving
Kyungmin Bin, Seungbeom Choi, Jimyoung Son, Jieun Choi, Daseul Bae, Daehyeon Baek, Kihyo Moon, Minsung Jang, Hyojung Lee
https://arxiv.org/abs/2509.06261

FineServe: Precision-Aware KV Slab and Two-Level Scheduling for Heterogeneous Precision LLM Serving
Recent advances in Post-Training Quantization (PTQ) techniques have significantly increased demand for serving quantized large language models (LLMs), enabling higher throughput and substantially reduced memory usage with minimal accuracy loss. Quantized models address memory constraints in LLMs and enhance GPU resource utilization through efficient GPU sharing. However, quantized models have smaller KV block sizes than non-quantized models, causing limited memory efficiency due to memory fragm…

Dynamical Quantum Multigraphs
Kassahun H. Betre, Nathan Lewis
https://arxiv.org/abs/2509.08296 https://arxiv.org/pdf/2509.08296

Dynamical Quantum Multigraphs
Motivated by applications in background-independent quantum gravity, we discuss the quantization of labeled and unlabeled finite multigraphs with a maximum edge count. We provide a unified way to represent quantum multigraphs with labeled or unlabeled vertices, which enables the study of quantum multigraphs as dynamical microscopic degrees of freedom and not just as representations of relations among quantum states of particles. The quantum multigraphs represent a quantum mechanical treatment o…

Crosslisted article(s) found for physics.optics. https://arxiv.org/list/physics.optics/new
[1/1]:
- Quantization of the electromagnetic fields from single atomic or molecular radiators
Valerica Raicu

A Theoretically-Grounded Codebook for Digital Semantic Communications
Lingyi Wang, Rashed Shelim, Walid Saad, Naren Ramakrishnan
https://arxiv.org/abs/2510.07108 https://…

A Theoretically-Grounded Codebook for Digital Semantic Communications
The use of a learnable codebook provides an efficient way for semantic communications to map vector-based high-dimensional semantic features onto discrete symbol representations required in digital communication systems. In this paper, the problem of codebook-enabled quantization mapping for digital semantic communications is studied from the perspective of information theory. Particularly, a novel theoretically-grounded codebook design is proposed for jointly optimizing quantization efficiency…

Deformation quantization of a hessian KV- structure on $\mathbb{R}^2$
Herguey Mopeng, Prosper Rosaire Mama Assandje, Joseph Dongho, Armand Tsimi
https://arxiv.org/abs/2509.23228

Deformation quantization of a hessian KV- structure on $\mathbb{R}^2$
This paper studies the quantization of the deformation of Hessian structures on a two-dimensional vector space, in the framework of Koszul-Vinberg algebras. We analyze how Hessian structures can be deformed to obtain quantum structures while preserving certain geometric and algebraic properties. The aim of the paper is to establish links between deformation theory and Hessian geometry.

Replaced article(s) found for cs.AR. https://arxiv.org/list/cs.AR/new
[1/1]:
- KLLM: Fast LLM Inference with K-Means Quantization
Xueying Wu, Baijun Zhou, Zhihui Gao, Yuzhe Fu, Qilin Zheng, Yintao He, Hai Li

Cat: Post-training quantization error reduction via cluster-based affine transformation
Ali Zoljodi, Radu Timofte, Masoud Daneshtalab
https://arxiv.org/abs/2509.26277 https://…

Cat: Post-training quantization error reduction via cluster-based affine transformation
Post-Training Quantization (PTQ) reduces the memory footprint and computational overhead of deep neural networks by converting full-precision (FP) values into quantized and compressed data types. While PTQ is more cost-efficient than Quantization-Aware Training (QAT), it is highly susceptible to accuracy degradation under a low-bit quantization (LQ) regime (e.g., 2-bit). Affine transformation is a classical technique used to reduce the discrepancy between the information processed by a quantize…

From Faraday and Maxwell to Quantum Physics. The later story of the Electromagnetic Vector Potential
Tuck Choy, Miguel Ortuno
https://arxiv.org/abs/2509.04486 https://

From Faraday and Maxwell to Quantum Physics. The later story of the Electromagnetic Vector Potential
With the advent of quantum mechanics by Heisenberg in 1925 exactly a century ago, the quantization of the electromagnetic field became an important goal for our founding fathers, whom we are here to celebrate. It was realized very soon that a consistent picture of quantum electrodynamics (QED) requires the quantization of not just the electromagnetic field ${\bf A}({\bf r},t)$, but the electron field ${\bf ψ}({\bf r},t)$ as well. The electron field is now a Dirac spinor field and it becomes a …

Crosslisted article(s) found for physics.chem-ph. https://arxiv.org/list/physics.chem-ph/new
[1/1]:
- Recursive algorithm for constructing antisymmetric fermionic states in first quantization mapping
E. Rule, I. A. Chernyshev, I. Stetcu, J. Carlson, R. Weiss

Mixture of Many Zero-Compute Experts: A High-Rate Quantization Theory Perspective
Yehuda Dar
https://arxiv.org/abs/2510.03151 https://arxiv.org/pdf/2510.03…

Mixture of Many Zero-Compute Experts: A High-Rate Quantization Theory Perspective
This paper uses classical high-rate quantization theory to provide new insights into mixture-of-experts (MoE) models for regression tasks. Our MoE is defined by a segmentation of the input space to regions, each with a single-parameter expert that acts as a constant predictor with zero-compute at inference. Motivated by high-rate quantization theory assumptions, we assume that the number of experts is sufficiently large to make their input-space regions very small. This lets us to study the app…

Progressive Semantic Residual Quantization for Multimodal-Joint Interest Modeling in Music Recommendation
Shijia Wang, Tianpei Ouyang, Qiang Xiao, Dongjing Wang, Yintao Ren, Songpei Xu, Da Guo, Chuanjiang Luo
https://arxiv.org/abs/2508.20359

Progressive Semantic Residual Quantization for Multimodal-Joint Interest Modeling in Music Recommendation
In music recommendation systems, multimodal interest learning is pivotal, which allows the model to capture nuanced preferences, including textual elements such as lyrics and various musical attributes such as different instruments and melodies. Recently, methods that incorporate multimodal content features through semantic IDs have achieved promising results. However, existing methods suffer from two critical limitations: 1) intra-modal semantic degradation, where residual-based quantization p…

Distributed Platoon Control Under Quantization: Stability Analysis and Privacy Preservation
Kaixiang Zhang, Zhaojian Li, Wei Lin
https://arxiv.org/abs/2510.05959 https://…

Distributed Platoon Control Under Quantization: Stability Analysis and Privacy Preservation
Distributed control of connected and automated vehicles has attracted considerable interest for its potential to improve traffic efficiency and safety. However, such control schemes require sharing privacy-sensitive vehicle data, which introduces risks of information leakage and potential malicious activities. This paper investigates the stability and privacy-preserving properties of distributed platoon control under two types of quantizers: deterministic and probabilistic. For deterministic qu…

AQUA-LLM: Evaluating Accuracy, Quantization, and Adversarial Robustness Trade-offs in LLMs for Cybersecurity Question Answering
Onat Gungor, Roshan Sood, Harold Wang, Tajana Rosing
https://arxiv.org/abs/2509.13514

AQUA-LLM: Evaluating Accuracy, Quantization, and Adversarial Robustness Trade-offs in LLMs for Cybersecurity Question Answering
Large Language Models (LLMs) have recently demonstrated strong potential for cybersecurity question answering (QA), supporting decision-making in real-time threat detection and response workflows. However, their substantial computational demands pose significant challenges for deployment on resource-constrained edge devices. Quantization, a widely adopted model compression technique, can alleviate these constraints. Nevertheless, quantization may degrade model accuracy and increase susceptibili…

Batalin-Fradkin-Vilkovisky Quantization of FLPR model
Ansha S. Nair, Saurabh Gupta
https://arxiv.org/abs/2509.05632 https://arxiv.org/pdf/2509.05632…

Batalin-Fradkin-Vilkovisky Quantization of FLPR model
We quantize the Friedberg-Lee-Pang-Ren (FLPR) model within the framework of Batalin-Fradkin-Vilkovisky (BFV) formalism. We construct the nilpotent Becchi-Rouet-Stora-Tyutin (BRST) charges using constraints and the fermionic gauge-fixing function by means of admissible gauge conditions. We also derive the BRST invariant effective action and corresponding symmetries of the model in both polar and Cartesian coordinates. Additionally, with the help of BRST charge, we demonstrate that the physical s…

A quantization of the $\operatorname{SL}_2(\mathbb{C})$-Chern-Simons invariant of tangle exteriors
Calvin McPhail-Snyder
https://arxiv.org/abs/2509.02365 https://

A quantization of the $\operatorname{SL}_2(\mathbb{C})$-Chern-Simons invariant of tangle exteriors
Previous work of the author and Reshetikhin shows how to use modules over the unrestricted quantum group (at a root of unity) to define quantum representations of a groupoid of braids with flat $\mathfrak{sl}_{2}$ connections on their complements, equivalently $\operatorname{SL}_{2}(\mathbb{C})$ representations or hyperbolic structures. In this article we explain how to use these representations to define quantum invariants $\mathcal{Z}_{N}$ of tangles and links with such geometric structures. …

Studying $\textrm{QED}_3$ with radial quantization on the lattice -- I. Free limit
Peter A. Boyle, Richard C. Brower, George T. Fleming, Emanuel Katz, Nobuyuki Matsumoto, Rohan Misra
https://arxiv.org/abs/2510.03085

Studying $\textrm{QED}_3$ with radial quantization on the lattice -- I. Free limit
To investigate the three-dimensional quantum electrodynamics in the radial quantization on the lattice, the lattice action is constructed and the free limit is studied on $S^2 \times \mathbb{R}$. With the overlap fermion, it is numerically verified that the important symmetries of the theory can be realized on the lattice. The analytic correlators are derived and compared to the lattice results, which agree including the overall normalization. The $O(a^2)$-scaling is confirmed toward the analyt…

Replaced article(s) found for cond-mat.str-el. https://arxiv.org/list/cond-mat.str-el/new
[1/1]:
- Exact quantization of topological order parameter in SU($N$) spin models, $N$-ality transformatio...
Hang Su, Yuan Yao, Akira Furusaki

How Quantization Shapes Bias in Large Language Models
Federico Marcuzzi, Xuefei Ning, Roy Schwartz, Iryna Gurevych
https://arxiv.org/abs/2508.18088 https://

How Quantization Shapes Bias in Large Language Models
This work presents a comprehensive evaluation of how quantization affects model bias, with particular attention to its impact on individual demographic subgroups. We focus on weight and activation quantization strategies and examine their effects across a broad range of bias types, including stereotypes, toxicity, sentiment, and fairness. We employ both probabilistic and generated text-based metrics across nine benchmarks and evaluate models varying in architecture family and reasoning ability.…

$C^0$-rigidity of Legendrians and coisotropics via sheaf quantization
Tomohiro Asano, Yuichi Ike, Christopher Kuo, Wenyuan Li
https://arxiv.org/abs/2510.01746 https://

$C^0$-rigidity of Legendrians and coisotropics via sheaf quantization
We prove that in the standard cosphere bundle, for any contact homeomorphism in the closure of the compactly supported contactomorphism group, when the image of a coisotropic submanifold (not necessarily properly embedded) is smooth, it is still coisotropic. Moreover, when contactomorphisms in the sequence are in the identity component and the image of a Legendrian is smooth, the Maslov data is preserved, and the category of sheaves with singular support on the Legendrian and the microstalk cor…

DiVeQ: Differentiable Vector Quantization Using the Reparameterization Trick
Mohammad Hassan Vali, Tom B\"ackstr\"om, Arno Solin
https://arxiv.org/abs/2509.26469 https…

DiVeQ: Differentiable Vector Quantization Using the Reparameterization Trick
Vector quantization is common in deep models, yet its hard assignments block gradients and hinder end-to-end training. We propose DiVeQ, which treats quantization as adding an error vector that mimics the quantization distortion, keeping the forward pass hard while letting gradients flow. We also present a space-filling variant (SF-DiVeQ) that assigns to a curve constructed by the lines connecting codewords, resulting in less quantization error and full codebook usage. Both methods train end-to…

Quantization of blow-up masses for the Finsler $N$-Liouville equation
Xia Huang, Yuan Li, Dong Ye, Feng Zhou
https://arxiv.org/abs/2508.16080 https://arxiv…

Quantization of blow-up masses for the Finsler $N$-Liouville equation
The quantization results for blow-up phenomena play crucial roles in the analysis of partial differential equations. Here we quantify the blow-up masses to the following Finsler $N$-Liouville equation $$-Q_{N}u_{n}=V_{n}e^{u_{n}}\quad\mbox{in}~ Ω\subset \mathbb{R}^{N}, N \ge 2.$$ Our study generalizes the classical result of Li-Shafrir [Indiana Univ. Math.J.,1994] for Liouville equation, Wang-Xia's work for anisotropic Liouville equation in $\mathbb{R}^2$ [JDE, 2012], and Esposito-Lucia's for …

Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Models
Donghoon Kim, Dongyoung Lee, Ik Joon Chang, Sung-Ho Bae
https://arxiv.org/abs/2509.26436

Post-Training Quantization via Residual Truncation and Zero Suppression for Diffusion Models
Diffusion models achieve high-quality image generation but face deployment challenges due to their high computational requirements. Although 8-bit outlier-aware post-training quantization (PTQ) matches full-precision performance, extending PTQ to 4 bits remains challenging. Larger step sizes in 4-bit quantization amplify rounding errors in dense, low-magnitude activations, leading to the loss of fine-grained textures. We hypothesize that not only outliers but also small activations are critical…

Word2Spike: Poisson Rate Coding for Associative Memories and Neuromorphic Algorithms
Archit Kalra, Midhun Sadanand
https://arxiv.org/abs/2509.07361 https://

Word2Spike: Poisson Rate Coding for Associative Memories and Neuromorphic Algorithms
Spiking neural networks offer a promising path toward energy-efficient, brain-like associative memory. This paper introduces Word2Spike, a novel rate coding mechanism that combines continuous word embeddings and neuromorphic architectures. We develop a one-to-one mapping that converts multi-dimensional word vectors into spike-based attractor states using Poisson processes. Using BitNet b1.58 quantization, we maintain 97% semantic similarity of continuous embeddings on SimLex-999 while achieving…

Non-Lagrangian Construction of Anyons via Flux Quantization in Cohomotopy
Hisham Sati, Urs Schreiber
https://arxiv.org/abs/2509.02577 https://arxiv.org/pdf…

Non-Lagrangian Construction of Anyons via Flux Quantization in Cohomotopy
We provide a brief invitation to the novel understanding of anyonic topological order in fractional quantum (anomalous) Hall systems, via "extraordinary" quantization of effective magnetic flux in Cohomotopy -- following our presentation at ISQS29.

Classical Polymerization of the Bianchi I Model with Deformed Poisson Structure
Babak Vakili
https://arxiv.org/abs/2510.06628 https://arxiv.org/pdf/2510.06…

Classical Polymerization of the Bianchi I Model with Deformed Poisson Structure
We study the dynamics of the Bianchi~I cosmological model in the presence of both polymer quantization effects and an exponential deformation of the Poisson algebra. Starting from the Hamiltonian formulation, we derive the polymer-deformed equations of motion and analyze their solutions for the contracting branch of the model. In contrast with the undeformed classical dynamics, the exponential deformation with suitable values of deformation parameters, produces a noticeably slower evolution of …

Using Landau quantization to probe disorder in semiconductor heterostructures
Asser Elsayed, Davide Costa, Lucas E. A. Stehouwer, Alberto Tosato, Mario Lodari, Brian Paquelet Wuetz, Davide Degli Esposti, Giordano Scappucci
https://arxiv.org/abs/2510.02794

Using Landau quantization to probe disorder in semiconductor heterostructures
Understanding scattering mechanisms in semiconductor heterostructures is crucial to reducing sources of disorder and ensuring high yield and uniformity in large spin qubit arrays. Disorder of the parent two-dimensional electron or hole gas is commonly estimated by the critical, percolation-driven density associated with the metal-insulator transition. However, a reliable estimation of the critical density within percolation theory is hindered by the need to measure conductivity with high precis…

LiquidGEMM: Hardware-Efficient W4A8 GEMM Kernel for High-Performance LLM Serving
Huanqi Hu, Bowen Xiao, Shixuan Sun, Jianian Yin, Zhexi Zhang, Xiang Luo, Chengquan Jiang, Weiqi Xu, Xiaoying Jia, Xin Liu, Minyi Guo
https://arxiv.org/abs/2509.01229

LiquidGEMM: Hardware-Efficient W4A8 GEMM Kernel for High-Performance LLM Serving
Quantization is a critical technique for accelerating LLM inference by reducing memory footprint and improving computational efficiency. Among various schemes, 4-bit weight and 8-bit activation quantization (W4A8) offers a strong balance between accuracy and performance. However, existing W4A8 GEMM kernels fall short in practice due to inefficient dequantization on CUDA Cores, which cannot keep pace with the high throughput of Tensor Cores. In this paper, we present LiquidGEMM, a hardware-effic…

Binary Weight Multi-Bit Activation Quantization for Compute-in-Memory CNN Accelerators
Wenyong Zhou, Zhengwu Liu, Yuan Ren, Ngai Wong
https://arxiv.org/abs/2508.21524 https://…

Binary Weight Multi-Bit Activation Quantization for Compute-in-Memory CNN Accelerators
Compute-in-memory (CIM) accelerators have emerged as a promising way for enhancing the energy efficiency of convolutional neural networks (CNNs). Deploying CNNs on CIM platforms generally requires quantization of network weights and activations to meet hardware constraints. However, existing approaches either prioritize hardware efficiency with binary weight and activation quantization at the cost of accuracy, or utilize multi-bit weights and activations for greater accuracy but limited efficie…

Distributed Detection and Bandwidth Allocation with Hybrid Quantized and Full-Precision Observations over Multiplicative Fading Channels
Linlin Mao, Zeping Sui, Michail Matthaiou, Hongbin Li
https://arxiv.org/abs/2510.06429

Distributed Detection and Bandwidth Allocation with Hybrid Quantized and Full-Precision Observations over Multiplicative Fading Channels
A hybrid detector that fuses both quantized and full-precision observations is proposed for weak signal detection under additive and multiplicative Gaussian noise. We first derive a locally most powerful test (LMPT)--based hybrid detector from the composite probability distribution of the compound observations received by the fusion center, and then analyze its asymptotic detection performance. Subsequently, we optimize the sensor-wise quantization thresholds to achieve near-optimal asymptotic …

AMAQ: Adaptive Mixed-bit Activation Quantization for Collaborative Parameter Efficient Fine-tuning
Yurun Song, Zhuoyi Yang, Ian G. Harris, Sangeetha Abdu Jyothi
https://arxiv.org/abs/2510.05468

AMAQ: Adaptive Mixed-bit Activation Quantization for Collaborative Parameter Efficient Fine-tuning
Large Language Models (LLMs) are scaling rapidly, creating significant challenges for collaborative server client distributed training, particularly in terms of communication efficiency and computational overheads. To address these challenges, we implement Parameter-efficient Split Learning, which effectively balances efficiency and performance for collaborative training on low-resource devices. To reduce communication overhead in collaborative training, we introduce Adaptive Mixed bit Activa…

Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs
Haokun Lin, Haobo Xu, Yichen Wu, Ziyu Guo, Renrui Zhang, Zhichao Lu, Ying Wei, Qingfu Zhang, Zhenan Sun
https://arxiv.org/abs/2508.14896

Quantization Meets dLLMs: A Systematic Study of Post-training Quantization for Diffusion LLMs
Recent advances in diffusion large language models (dLLMs) have introduced a promising alternative to autoregressive (AR) LLMs for natural language generation tasks, leveraging full attention and denoising-based decoding strategies. However, the deployment of these models on edge devices remains challenging due to their massive parameter scale and high resource demands. While post-training quantization (PTQ) has emerged as a widely adopted technique for compressing AR LLMs, its applicability to…

Spectrogram Patch Codec: A 2D Block-Quantized VQ-VAE and HiFi-GAN for Neural Speech Coding
Luis Felipe Chary, Miguel Arjona Ramirez
https://arxiv.org/abs/2509.02244 https://

Spectrogram Patch Codec: A 2D Block-Quantized VQ-VAE and HiFi-GAN for Neural Speech Coding
We present a neural speech codec that challenges the need for complex residual vector quantization (RVQ) stacks by introducing a simpler, single-stage quantization approach. Our method operates directly on the mel-spectrogram, treating it as a 2D data and quantizing non-overlapping 4x4 patches into a single, shared codebook. This patchwise design simplifies the architecture, enables low-latency streaming, and yields a discrete latent grid. To ensure high-fidelity synthesis, we employ a late-sta…

Operator Algebras and Third Quantization
Yidong Chen, Marius Junge, Nima Lashkari
https://arxiv.org/abs/2509.02293 https://arxiv.org/pdf/2509.02293

Operator Algebras and Third Quantization
In quantum gravity, the gravitational path integral involves a sum over topologies, representing the joining and splitting of multiple universes. To account for topology change, one is led to allow the creation and annihilation of closed and open universes in a framework often called third quantization or universe field theory. We argue that since topology change in gravity is a rare event, its contribution to exponentially late-time physics is universally described by a Poisson process. This u…

DPQuant: Efficient and Differentially-Private Model Training via Dynamic Quantization Scheduling
Yubo Gao, Renbo Tu, Gennady Pekhimenko, Nandita Vijaykumar
https://arxiv.org/abs/2509.03472

DPQuant: Efficient and Differentially-Private Model Training via Dynamic Quantization Scheduling
Differentially-Private SGD (DP-SGD) is a powerful technique to protect user privacy when using sensitive data to train neural networks. During training, converting model weights and activations into low-precision formats, i.e., quantization, can drastically reduce training times, energy consumption, and cost, and is thus a widely used technique. In this work, we demonstrate that quantization causes significantly higher accuracy degradation in DP-SGD compared to regular SGD. We observe that this…

A Node-Aware Dynamic Quantization Approach for Graph Collaborative Filtering
Lin Li, Chunyang Li, Yu Yin, Xiaohui Tao, Jianwei Zhang
https://arxiv.org/abs/2508.16516 https://

A Node-Aware Dynamic Quantization Approach for Graph Collaborative Filtering
In the realm of collaborative filtering recommendation systems, Graph Neural Networks (GNNs) have demonstrated remarkable performance but face significant challenges in deployment on resource-constrained edge devices due to their high embedding parameter requirements and computational costs. Using common quantization method directly on node embeddings may overlooks their graph based structure, causing error accumulation during message passing and degrading the quality of quantized embeddings.To…

Analysis of polymerized superconducting circuits
Sean Crowe, Stefan Evans, Alexei Smolyaninov
https://arxiv.org/abs/2509.18016 https://arxiv.org/pdf/2509.1…

Analysis of polymerized superconducting circuits
We apply polymer quantization, a quantization technique sometimes used in high energy physics, to several superconducting circuits including: transmons, transmission line resonators, and LC circuits. In the case of transmon qubits and transmission line resonators, experimental predictions are very close to what is found with canonical quantization, though in this approach constant charge offsets can be interpreted as quantization ambiguities. In the case of LC circuits, polymer quantization pre…

Systematic Characterization of LLM Quantization: A Performance, Energy, and Quality Perspective
Tianyao Shi, Yi Ding
https://arxiv.org/abs/2508.16712 https://

Systematic Characterization of LLM Quantization: A Performance, Energy, and Quality Perspective
Large language models (LLMs) have demonstrated remarkable capabilities across diverse domains, but their heavy resource demands make quantization-reducing precision to lower-bit formats-critical for efficient serving. While many quantization methods exist, a systematic understanding of their performance, energy, and quality tradeoffs in realistic serving conditions remains a gap. In this work, we first develop a fully automated online characterization framework qMeter, and then conduct an in-de…

Stratification of the half-density quantization of the Jeffrey-Weitsman-Witten invariants
Adrian Chitan
https://arxiv.org/abs/2509.17656 https://arxiv.org/…

Stratification of the half-density quantization of the Jeffrey-Weitsman-Witten invariants
In this paper we review the quantization procedure employed by Jeffrey and Weitsman in their formalization of Witten's three manifold invariants. In so doing, we formalize the stratification of the Lagrangian leaf which corresponds to the flat connections of a bounding surface which extend into the handlebody. The smooth structure of this space motivates a procedure for a stratified quantization: use volumes and corresponding half-densities on the stratified tangent bundle. The pairing is then …

Continuum Fractons: Quantization and the Many Body Problem
Ylias Sadki, Abhishodh Prakash, S. L. Sondhi
https://arxiv.org/abs/2510.00110 https://arxiv.org/…

Continuum Fractons: Quantization and the Many Body Problem
We formulate a continuum quantum mechanics for non-relativistic, dipole-conserving fractons. Imposing symmetries and locality results in novel phenomena absent in ordinary quantum mechanical systems. A single fracton has a vanishing Hamiltonian, and thus its spectrum is entirely composed of zero modes. For the two-body problem, the Hamiltonian is perfectly described by Sturm--Liouville (SL) theory. The effective two-body Hamiltonian is an SL operator on $(-1,1)$ whose spectral type is set by th…

Thermodynamics of Kerr black hole: Tsallis-Cirto composition law and entropy quantization
G. E. Volovik
https://arxiv.org/abs/2509.00748 https://arxiv.org/…

Thermodynamics of Kerr black hole: Tsallis-Cirto composition law and entropy quantization
The processes of splitting and merging of black holes obey the composition law generated by the Tsallis-Cirto $δ=2$ statistics. The same composition law expresses the full entropy of the Reissner-Nordström black hole via the entropies of its outer and inner horizons. He we apply this composition law to the entropy of the Kerr black hole. As distinct from Reissner-Nordström black hole, where the full entropy depends only on mass $M$ and does not depend on its charge $Q$, the entropy of Kerr b…

Coproducts for affine super-Yangian and Weyl groupoid action
Vasiliy Volkov, Vladimir Stukopin
https://arxiv.org/abs/2510.04221 https://arxiv.org/pdf/2510.…

Coproducts for affine super-Yangian and Weyl groupoid action
For affine special linear superalgebra $\widehat{sl}(m|n, Π)$ defined by an arbitrary system of simple roots $Π$ we define the affine super Yangian $Y_{\hbar}(\widehat{sl}(m|n, Π))$ as Hopf superalgebra which is a quantization of superbialgebra $\widehat{sl}(m|n, Π)[t]$ and describe super Yangian in terms of minimalistic system of generators. We consider Drinfeld presentation for $Y^D_{\hbar}(\widehat{sl}(m|n, Π))$ and prove that these two presentations are isomorphic a…

Efficient Quantization-Aware Neural Receivers: Beyond Post-Training Quantization
SaiKrishna Saketh Yellapragada, Esa Ollila, Mario Costa
https://arxiv.org/abs/2509.13786 https:/…

Efficient Quantization-Aware Neural Receivers: Beyond Post-Training Quantization
As wireless communication systems advance toward Sixth Generation (6G) Radio Access Networks (RAN), Deep Learning (DL)-based neural receivers are emerging as transformative solutions for Physical Layer (PHY) processing, delivering superior Block Error Rate (BLER) performance compared to traditional model-based approaches. Practical deployment on resource-constrained hardware, however, requires efficient quantization to reduce latency, energy, and memory without sacrificing reliability. We exten…

TinyMusician: On-Device Music Generation with Knowledge Distillation and Mixed Precision Quantization
Hainan Wang, Mehdi Hosseinzadeh, Reza Rawassizadeh
https://arxiv.org/abs/2509.00914

TinyMusician: On-Device Music Generation with Knowledge Distillation and Mixed Precision Quantization
The success of the generative model has gained unprecedented attention in the music generation area. Transformer-based architectures have set new benchmarks for model performance. However, their practical adoption is hindered by some critical challenges: the demand for massive computational resources and inference time, due to their large number of parameters. These obstacles make them infeasible to deploy on edge devices, such as smartphones and wearables, with limited computational resources.…

Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer
Ziyuan Huang, DanDan Zheng, Cheng Zou, Rui Liu, Xiaolong Wang, Kaixiang Ji, Weilong Chai, Jianxin Sun, Libin Wang, Yongjie Lv, Taozhi Huang, Jiajia Liu, Qingpei Guo, Ming Yang, Jingdong Chen, Jun Zhou
https://arxiv.org/abs/2510.06590

Ming-UniVision: Joint Image Understanding and Generation with a Unified Continuous Tokenizer
Visual tokenization remains a core challenge in unifying visual understanding and generation within the autoregressive paradigm. Existing methods typically employ tokenizers in discrete latent spaces to align with the tokens from large language models, where the quantization errors can limit semantic expressiveness and degrade the capability of vision-language understanding. To address this, we introduce MingTok, a new family of visual tokenizers with a continuous latent space, for unified auto…

Phonological Representation Learning for Isolated Signs Improves Out-of-Vocabulary Generalization
Lee Kezar, Zed Sehyr, Jesse Thomason
https://arxiv.org/abs/2509.04745 https://

Phonological Representation Learning for Isolated Signs Improves Out-of-Vocabulary Generalization
Sign language datasets are often not representative in terms of vocabulary, underscoring the need for models that generalize to unseen signs. Vector quantization is a promising approach for learning discrete, token-like representations, but it has not been evaluated whether the learned units capture spurious correlations that hinder out-of-vocabulary performance. This work investigates two phonological inductive biases: Parameter Disentanglement, an architectural bias, and Phonological Semi-Sup…

Bohr-Sommerfeld quantization conditions for Schrodinger operator: the Method of Microlocal Wronskian and Gram Matrix
Abdelwaheb Ifa, Michel Rouleux
https://arxiv.org/abs/2509.23514

Bohr-Sommerfeld quantization conditions for Schrodinger operator: the Method of Microlocal Wronskian and Gram Matrix
In this paper, we present a new formulation of the well known Bohr-Sommerfeld quantization rule (BS) of order 2 for a Schrodinger operator within the algebraic and microlocal framework of B. Helffer and J. Sjostrand; BS holds precisely when the Gram matrix consisting of scalar products of some WKB solutions with respect to the "flux norm" is not invertible. This condition is obtained using the microlocal Wronskian and does not rely on traditional matching techniques. It is simplified by using a…

Simultaneous Quantization and Reduction of Constrained Systems
Jianhao M. Yang
https://arxiv.org/abs/2509.22747 https://arxiv.org/pdf/2509.22747

Simultaneous Quantization and Reduction of Constrained Systems
We present a novel framework for quantizing constrained quantum systems in which the processes of quantization and constraint enforcement are performed simultaneously. The approach is based on an extension of the stationary action principle, incorporating an information-theoretic term arising from vacuum fluctuations. Constraints are included directly in the Lagrangian via Lagrange multipliers, allowing the subsequent variational procedure to yield the quantum dynamics without ambiguity regardi…

Enhancing Communication Efficiency in FL with Adaptive Gradient Quantization and Communication Frequency Optimization
Asadullah Tariq, Tariq Qayyum, Mohamed Adel Serhani, Farag Sallabi, Ikbal Taleb, Ezedin S. Barka
https://arxiv.org/abs/2509.23419

Enhancing Communication Efficiency in FL with Adaptive Gradient Quantization and Communication Frequency Optimization
Federated Learning (FL) enables participant devices to collaboratively train deep learning models without sharing their data with the server or other devices, effectively addressing data privacy and computational concerns. However, FL faces a major bottleneck due to high communication overhead from frequent model updates between devices and the server, limiting deployment in resource-constrained wireless networks. In this paper, we propose a three-fold strategy. Firstly, an Adaptive Feature-Eli…

All-optical band structure reconstruction and onset of Landau quantization of Dirac fermions
Josef Riepl (Department of Physics, University of Regensburg, Regensburg, Germany), Marc Aichner (Department of Physics, University of Regensburg, Regensburg, Germany, LSI, CNRS, CEA/DRF/IRAMIS, Ecole Polytechnique, Institut Polytechnique de Paris, Palaiseau, France), Nikolai N. Mikhailov (A.V. Rzhanov Institute of Semiconductor Physics, Novosibirsk, Russia), Sergei A. Dvoretsky (A.V. Rzhanov I…

All-optical band structure reconstruction and onset of Landau quantization of Dirac fermions
The nature of relativistic electrons in solids depends on the precise shape of the underlying band structure. Prominently, symmetry-related mechanisms, such as the breaking of time reversal symmetry in topological insulators, can lead to the emergence of band gaps on small energy scales. It is, thus, important to quantify potential gaps of the Dirac cone with meV precision. Yet established band structure measurements are often challenged by their strict surface sensitivity or limited energy res…

Can Less Precise Be More Reliable? A Systematic Evaluation of Quantization's Impact on CLIP Beyond Accuracy
Aymen Bouguerra, Daniel Montoya, Alexandra Gomez-Villa, Fabio Arnez, Chokri Mraidha
https://arxiv.org/abs/2509.21173

Can Less Precise Be More Reliable? A Systematic Evaluation of Quantization's Impact on CLIP Beyond Accuracy
The powerful zero-shot generalization capabilities of vision-language models (VLMs) like CLIP have enabled new paradigms for safety-related tasks such as out-of-distribution (OOD) detection. However, additional aspects crucial for the computationally efficient and reliable deployment of CLIP are still overlooked. In particular, the impact of quantization on CLIP's performance beyond accuracy remains underexplored. This work presents a large-scale evaluation of quantization on CLIP models, asses…

Deep Multiple Quantization Network on Long Behavior Sequence for Click-Through Rate Prediction
Zhuoxing Wei, Qi Liu, Qingchen Xie
https://arxiv.org/abs/2508.20865 https://

Deep Multiple Quantization Network on Long Behavior Sequence for Click-Through Rate Prediction
In Click-Through Rate (CTR) prediction, the long behavior sequence, comprising the user's long period of historical interactions with items has a vital influence on assessing the user's interest in the candidate item. Existing approaches strike efficiency and effectiveness through a two-stage paradigm: first retrieving hundreds of candidate-related items and then extracting interest intensity vector through target attention. However, we argue that the discrepancy in target attention's relevance…

Stochastic dynamics for group field theories II: Methods for nonequilibrium renormalization group
Vincent Lahoche, Dine Ousmane Samary
https://arxiv.org/abs/2509.05507 https://

Stochastic dynamics for group field theories II: Methods for nonequilibrium renormalization group
This paper is a continuation of our earlier work, which aimed to develop methods for understanding the renormalization group of tensorial group field theories within the stochastic quantization framework. In that first study, we showed that the equations governing melonic structures, together with Ward identities, make it possible to close the hierarchy of flow equations, thereby reproducing the results of equilibrium theory. In the present work, we go further by extending the formalism to the …

Beat-Based Rhythm Quantization of MIDI Performances
Maximilian Wachter, Sebastian Murgul, Michael Heizmann
https://arxiv.org/abs/2508.19262 https://arxiv.o…

Beat-Based Rhythm Quantization of MIDI Performances
We propose a transformer-based rhythm quantization model that incorporates beat and downbeat information to quantize MIDI performances into metrically-aligned, human-readable scores. We propose a beat-based preprocessing method that transfers score and performance data into a unified token representation. We optimize our model architecture and data representation and train on piano and guitar performances. Our model exceeds state-of-the-art performance based on the MUSTER metric.

Fair-GPTQ: Bias-Aware Quantization for Large Language Models
Irina Proskurina, Guillaume Metzler, Julien Velcin
https://arxiv.org/abs/2509.15206 https://ar…

Fair-GPTQ: Bias-Aware Quantization for Large Language Models
High memory demands of generative language models have drawn attention to quantization, which reduces computational cost, memory usage, and latency by mapping model weights to lower-precision integers. Approaches such as GPTQ effectively minimize input-weight product errors during quantization; however, recent empirical studies show that they can increase biased outputs and degrade performance on fairness benchmarks, and it remains unclear which specific weights cause this issue. In this work, …

AI-Driven Fronthaul Link Compression in Wireless Communication Systems: Review and Method Design
Keqin Zhang
https://arxiv.org/abs/2509.04805 https://arxiv…

AI-Driven Fronthaul Link Compression in Wireless Communication Systems: Review and Method Design
Modern fronthaul links in wireless systems must transport high-dimensional signals under stringent bandwidth and latency constraints, which makes compression indispensable. Traditional strategies such as compressed sensing, scalar quantization, and fixed-codec pipelines often rely on restrictive priors, degrade sharply at high compression ratios, and are hard to tune across channels and deployments. Recent progress in Artificial Intelligence (AI) has brought end-to-end learned transforms, vecto…

Optimal Remainder Estimates in the Quantization of Complex Projective Spaces
Tommaso Aschieri, B{\l}a\.zej Ruba, Jan Philip Solovej
https://arxiv.org/abs/2508.19968 https://

Optimal Remainder Estimates in the Quantization of Complex Projective Spaces
We study Berezin-Toeplitz quantization of complex projective spaces $\mathbb{CP}^{d-1}$ and obtain full asymptotic expansions of the Berezin transformation and of products of Toeplitz operators. In each case, the remainder is controlled by the next term of the expansion, either through a positivity-preserving transformation or via an operator inequality. This leads to bounds which are optimal in terms of the required regularity and feature sharp or asymptotically sharp constants.

SIRA: Scaled-Integer Range Analysis for Optimizing FPGA Dataflow Neural Network Accelerators
Yaman Umuroglu, Christoph Berganski, Felix Jentzsch, Michal Danilowicz, Tomasz Kryjak, Charalampos Bezaitis, Magnus Sjalander, Ian Colbert, Thomas Preusser, Jakoba Petri-Koenig, Michaela Blott
https://arxiv.org/abs/2508.21493

SIRA: Scaled-Integer Range Analysis for Optimizing FPGA Dataflow Neural Network Accelerators
While neural network quantization effectively reduces the cost of matrix multiplications, aggressive quantization can expose non-matrix-multiply operations as significant performance and resource bottlenecks on embedded systems. Addressing such bottlenecks requires a comprehensive approach to tailoring the precision across operations in the inference computation. To this end, we introduce scaled-integer range analysis (SIRA), a static analysis technique employing interval arithmetic to determin…

Scene-Aware Vectorized Memory Multi-Agent Framework with Cross-Modal Differentiated Quantization VLMs for Visually Impaired Assistance
Xiangxiang Wang, Xuanyu Wang, YiJia Luo, Yongbin Yu, Manping Fan, Jingtao Zhang, Liyong Ren
https://arxiv.org/abs/2508.18177

Scene-Aware Vectorized Memory Multi-Agent Framework with Cross-Modal Differentiated Quantization VLMs for Visually Impaired Assistance
This study proposes the dual technological innovation framework, including a cross-modal differ entiated quantization framework for vision-language models (VLMs) and a scene-aware vectorized memory multi-agent system for visually impaired assistance. The modular framework was developed implementing differentiated processing strategies, effectively reducing memory requirements from 38GB to 16GB while maintaining model performance. The multi-agent architecture combines scene classificatio…

Language Model Based Text-to-Audio Generation: Anti-Causally Aligned Collaborative Residual Transformers
Juncheng Wang, Chao Xu, Cheng Yu, Zhe Hu, Haoyu Xie, Guoqi Yu, Lei Shang, Shujun Wang
https://arxiv.org/abs/2510.04577

Language Model Based Text-to-Audio Generation: Anti-Causally Aligned Collaborative Residual Transformers
While language models (LMs) paired with residual vector quantization (RVQ) tokenizers have shown promise in text-to-audio (T2A) generation, they still lag behind diffusion-based models by a non-trivial margin. We identify a critical dilemma underpinning this gap: incorporating more RVQ layers improves audio reconstruction fidelity but exceeds the generation capacity of conventional LMs. To address this, we first analyze RVQ dynamics and uncover two key limitations: 1) orthogonality of features …

Renormalization of Chern-Simons Wilson Loops via Flux Quantization in Cohomotopy
Hisham Sati, Urs Schreiber
https://arxiv.org/abs/2509.25336 https://arxiv.…

Renormalization of Chern-Simons Wilson Loops via Flux Quantization in Cohomotopy
In the practice of physics model building, the process of renormalization, resummation, and anomaly cancellation is to incrementally repair initially ill-defined Lagrangian quantum field theories. Impressive as this is, one would rather have concisely defined complete theories to begin with, and understand these choices as emergent from fundamental principles. As an instructive example, we recall renormalization choices for Wilson loop observables in abelian Chern-Simons theory. Then we show th…

QeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMs
Wei Huang, Yi Ge, Shuai Yang, Yicheng Xiao, Huizi Mao, Yujun Lin, Hanrong Ye, Sifei Liu, Ka Chun Cheung, Hongxu Yin, Yao Lu, Xiaojuan Qi, Song Han, Yukang Chen
https://arxiv.org/abs/2510.11696

QeRL: Beyond Efficiency -- Quantization-enhanced Reinforcement Learning for LLMs
We propose QeRL, a Quantization-enhanced Reinforcement Learning framework for large language models (LLMs). While RL is essential for LLMs' reasoning capabilities, it is resource-intensive, requiring substantial GPU memory and long rollout durations. QeRL addresses these issues by combining NVFP4 quantization with Low-Rank Adaptation (LoRA), accelerating rollout phase of RL while reducing memory overhead. Beyond efficiency, our findings show that quantization noise increases policy entropy, enh…

$\gamma$-Quant: Towards Learnable Quantization for Low-bit Pattern Recognition
Mishal Fatima, Shashank Agnihotri, Marius Bock, Kanchana Vaishnavi Gandikota, Kristof Van Laerhoven, Michael Moeller, Margret Keuper
https://arxiv.org/abs/2509.22448

$γ$-Quant: Towards Learnable Quantization for Low-bit Pattern Recognition
Most pattern recognition models are developed on pre-proce\-ssed data. In computer vision, for instance, RGB images processed through image signal processing (ISP) pipelines designed to cater to human perception are the most frequent input to image analysis networks. However, many modern vision tasks operate without a human in the loop, raising the question of whether such pre-processing is optimal for automated analysis. Similarly, human activity recognition (HAR) on body-worn sensor data comm…

Exact WKB Formulation of Quantization and Particle Production in Time-Dependent Backgrounds
Ryo Namba, Motoo Suzuki
https://arxiv.org/abs/2509.19194 https://

LLM Compression: How Far Can We Go in Balancing Size and Performance?
Sahil Sk, Debasish Dhal, Sonal Khosla, Sk Shahid, Sambit Shekhar, Akash Dhaka, Shantipriya Parida, Dilip K. Prasad, Ond\v{r}ej Bojar
https://arxiv.org/abs/2508.11318

LLM Compression: How Far Can We Go in Balancing Size and Performance?
Quantization is an essential and popular technique for improving the accessibility of large language models (LLMs) by reducing memory usage and computational costs while maintaining performance. In this study, we apply 4-bit Group Scaling Quantization (GSQ) and Generative Pretrained Transformer Quantization (GPTQ) to LLaMA 1B, Qwen 0.5B, and PHI 1.5B, evaluating their impact across multiple NLP tasks. We benchmark these models on MS MARCO (Information Retrieval), BoolQ (Boolean Question Answeri…

AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models
Sangjun Lee, Seung-taek Woo, Jungyu Jin, Changhun Lee, Eunhyeok Park
https://arxiv.org/abs/2509.12019

AMQ: Enabling AutoML for Mixed-precision Weight-Only Quantization of Large Language Models
To enable broader deployment of Large Language Models (LLMs), it is essential to identify the best-performing model under strict memory constraints. We present AMQ, Automated Mixed-Precision Weight-Only Quantization, a framework that assigns layer-wise quantization bit-widths to optimally balance model quality and memory usage. However, the combinatorial search space, with over 10^{100} possible configurations, makes conventional black-box optimization infeasible. AMQ overcomes this challenge t…

On the Quantization of the Electromagnetic Field with Magnetic Monopoles
Kanan Anwar
https://arxiv.org/abs/2509.17284 https://arxiv.org/pdf/2509.17284

On the Quantization of the Electromagnetic Field with Magnetic Monopoles
We present a covariant framework for the quantization of the electromagnetic field in the presence of magnetic monopoles. Building on the two-potential formalism of Cabibbo and Ferrari, which treats electric and magnetic sources on equal footing and reveals a $U(1) \times U(1)$ gauge symmetry, we extend the theory into the quantum domain. Using the Gupta-Bleuler procedure in the Feynman-'t Hooft gauge, we construct the physical Hilbert space and eliminate negative-norm states. The resulting the…

Making Pose Representations More Expressive and Disentangled via Residual Vector Quantization
Sukhyun Jeong, Hong-Gi Shin, Yong-Hoon Choi
https://arxiv.org/abs/2508.14561 https:…

Making Pose Representations More Expressive and Disentangled via Residual Vector Quantization
Recent progress in text-to-motion has advanced both 3D human motion generation and text-based motion control. Controllable motion generation (CoMo), which enables intuitive control, typically relies on pose code representations, but discrete pose codes alone cannot capture fine-grained motion details, limiting expressiveness. To overcome this, we propose a method that augments pose code-based latent representations with continuous motion features using residual vector quantization (RVQ). This d…

Chunk Based Speech Pre-training with High Resolution Finite Scalar Quantization
Yun Tang, Cindy Tseng
https://arxiv.org/abs/2509.15579 https://arxiv.org/pd…

Chunk Based Speech Pre-training with High Resolution Finite Scalar Quantization
Low latency speech human-machine communication is becoming increasingly necessary as speech technology advances quickly in the last decade. One of the primary factors behind the advancement of speech technology is self-supervised learning. Most self-supervised learning algorithms are designed with full utterance assumption and compromises have to made if partial utterances are presented, which are common in the streaming applications. In this work, we propose a chunk based self-supervised learn…

CHORD: Customizing Hybrid-precision On-device Model for Sequential Recommendation with Device-cloud Collaboration
Tianqi Liu, Kairui Fu, Shengyu Zhang, Wenyan Fan, Zhaocheng Du, Jieming Zhu, Fan Wu, Fei Wu
https://arxiv.org/abs/2510.03038

CHORD: Customizing Hybrid-precision On-device Model for Sequential Recommendation with Device-cloud Collaboration
With the advancement of mobile device capabilities, deploying reranking models directly on devices has become feasible, enabling real-time contextual recommendations. When migrating models from cloud to devices, resource heterogeneity inevitably necessitates model compression. Recent quantization methods show promise for efficient deployment, yet they overlook device-specific user interests, resulting in compromised recommendation accuracy. While on-device finetuning captures personalized user …

GDNSQ: Gradual Differentiable Noise Scale Quantization for Low-bit Neural Networks
Sergey Salishev, Ian Akhremchik
https://arxiv.org/abs/2508.14004 https://

GDNSQ: Gradual Differentiable Noise Scale Quantization for Low-bit Neural Networks
Quantized neural networks can be viewed as a chain of noisy channels, where rounding in each layer reduces capacity as bit-width shrinks; the floating-point (FP) checkpoint sets the maximum input rate. We track capacity dynamics as the average bit-width decreases and identify resulting quantization bottlenecks by casting fine-tuning as a smooth, constrained optimization problem. Our approach employs a fully differentiable Straight-Through Estimator (STE) with learnable bit-width, noise scale an…

Exploiting Information Redundancy in Attention Maps for Extreme Quantization of Vision Transformers
Lucas Maisonnave, Karim Haroun, Tom Pegeot
https://arxiv.org/abs/2508.16311 h…

Exploiting Information Redundancy in Attention Maps for Extreme Quantization of Vision Transformers
Transformer models rely on Multi-Head Self-Attention (MHSA) mechanisms, where each attention head contributes to the final representation. However, their computational complexity and high memory demands due to MHSA hinders their deployment at the edge. In this work, we analyze and exploit information redundancy in attention maps to accelerate model inference. By quantifying the information captured by each attention head using Shannon entropy, our analysis reveals that attention heads with lowe…

Unstable mode and the Unruh-DeWitt detector
Bruno S. Felipe, Jo\~ao P. M. Pitelli
https://arxiv.org/abs/2508.20993 https://arxiv.org/pdf/2508.20993

Unstable mode and the Unruh-DeWitt detector
We investigate the quantization of a single unstable mode in a real scalar field subject to a Robin boundary condition in (1+1)-dimensional half-Minkowski spacetime. The instability arises from an imaginary frequency mode - analogous to that of the inverted harmonic oscillator - requiring the rigged Hilbert space formalism for consistent quantization. Within this framework, the unstable mode is naturally described as a well-defined decaying (or growing) quantum state with a characteristic mean …

Enhancing Model Privacy in Federated Learning with Random Masking and Quantization
Zhibo Xu, Jianhao Zhu, Jingwen Xu, Changze Lv, Zisu Huang, Xiaohua Wang, Muling Wu, Qi Qian, Xiaoqing Zheng, Xuanjing Huang
https://arxiv.org/abs/2508.18911

Enhancing Model Privacy in Federated Learning with Random Masking and Quantization
Experimental results across various models and tasks demonstrate that our approach not only maintains strong model performance in federated learning settings but also achieves enhanced protection of model parameters compared to baseline methods.

Image-Conditioned 3D Gaussian Splat Quantization
Xinshuang Liu, Runfa Blark Li, Keito Suzuki, Truong Nguyen
https://arxiv.org/abs/2508.15372 https://arxiv.…

Image-Conditioned 3D Gaussian Splat Quantization
3D Gaussian Splatting (3DGS) has attracted considerable attention for enabling high-quality real-time rendering. Although 3DGS compression methods have been proposed for deployment on storage-constrained devices, two limitations hinder archival use: (1) they compress medium-scale scenes only to the megabyte range, which remains impractical for large-scale scenes or extensive scene collections; and (2) they lack mechanisms to accommodate scene changes after long-term archival. To address these l…

Q-Palette: Fractional-Bit Quantizers Toward Optimal Bit Allocation for Efficient LLM Deployment
Deokjae Lee, Hyun Oh Song
https://arxiv.org/abs/2509.20214 https://

Q-Palette: Fractional-Bit Quantizers Toward Optimal Bit Allocation for Efficient LLM Deployment
We study weight-only post-training quantization (PTQ), which quantizes the weights of a large language model (LLM) without retraining, using little or no calibration data. Weight-only PTQ is crucial for reducing the memory footprint and latency of LLM inference, especially in memory-bound, small-batch inference scenarios, such as personalized inference on edge devices. Despite its importance, irregular weight distributions with heavy-tailed outliers in LLMs complicate quantization, recently mot…

CARVQ: Corrective Adaptor with Group Residual Vector Quantization for LLM Embedding Compression
Dayin Gou, Sanghyun Byun, Nilesh Malpeddi, Gabrielle De Micheli, Prathamesh Vaste, Jacob Song, Woo Seong Chung
https://arxiv.org/abs/2510.12721

CARVQ: Corrective Adaptor with Group Residual Vector Quantization for LLM Embedding Compression
Large Language Models (LLMs) typically rely on a large number of parameters for token embedding, leading to substantial storage requirements and memory footprints. In particular, LLMs deployed on edge devices are memory-bound, and reducing the memory footprint by compressing the embedding layer not only frees up the memory bandwidth but also speeds up inference. To address this, we introduce CARVQ, a post-training novel Corrective Adaptor combined with group Residual Vector Quantization. CARVQ …