arXiv subjects

TianQi Hou

Publications and source records attributed to TianQi Hou.

8 recordsLinked to original sources

arXiv2026

Diving into Kronecker Adapters: Component Design Matters

Kronecker adapters have emerged as a promising approach for fine-tuning large-scale models, enabling high-rank updates through tunable component structures. However, existing work largely treats the component structure as a fixed or heuristic design choice, leaving the dimensions and number of Kronecker components underexplored. In this paper, we identify component structure as a key factor governing the capacity of Kronecker adapters. We perform a fine-grained analysis of both the dimensions and number of Kronecker components. In particular, we show that the alignment between Kronecker adapters and full fine-tuning depends on component configurations. Guided by these insights, we propose Component Designed Kronecker Adapters (CDKA). We further provide parameter-budget-aware configuration guidelines and a tailored training stabilization strategy for practical deployment. Experiments across various architectures and modalities demonstrate the effectiveness of CDKA. Code is available at https://github.com/rainstonee/CDKA.

cs.LG↗

arXiv2026

On the Interpolation Error of Nonlinear Attention versus Linear Regression

Attention has become the core building block of modern machine learning (ML) by efficiently capturing the long-range dependencies among input tokens. Its inherently parallelizable structure allows for efficient performance scaling with the rapidly increasing size of both data and model parameters. Despite its central role, the theoretical understanding of Attention, especially in the nonlinear setting, is progressing at a more modest pace. This paper provides a precise characterization of the interpolation error for a nonlinear Attention, in the high-dimensional regime where the number of input tokens $n$ and the embedding dimension $p$ are both large and comparable. Under a signal-plus-noise data model and for fixed Attention weights, we derive explicit (limiting) expressions for the mean-squared interpolation error. Leveraging recent advances in random matrix theory, we show that nonlinear Attention generally incurs a larger interpolation error than linear regression on random inputs. However, this gap vanishes, and can even be reversed, when the input contains a structured signal, particularly if the Attention weights align with the signal direction. Our theoretical insights are supported by numerical experiments.

stat.ML↗

arXiv2023

Capacity-Achieving Sparse Regression Codes via Vector Approximate Message Passing

Sparse regression codes (SPARCs) are a promising coding scheme that can approach the Shannon limit over Additive White Gaussian Noise (AWGN) channels. Previous works have proven the capacity-achieving property of SPARCs with Gaussian design matrices. We generalize these results to right orthogonally invariant ensembles that allow for more structured design matrices. With the Vector Approximate Message Passing (VAMP) decoder, we rigorously demonstrate the exponentially decaying error probability for design matrices that satisfy a certain criterion with the exponentially decaying power allocation. For other spectra, we design a new power allocation scheme to show that the information theoretical threshold is achievable.

cs.IT↗

arXiv2023

Mismatched estimation of non-symmetric rank-one matrices corrupted by structured noise

We study the performance of a Bayesian statistician who estimates a rank-one signal corrupted by non-symmetric rotationally invariant noise with a generic distribution of singular values. As the signal-to-noise ratio and the noise structure are unknown, a Gaussian setup is incorrectly assumed. We derive the exact analytic expression for the error of the mismatched Bayes estimator and also provide the analysis of an approximate message passing (AMP) algorithm. The first result exploits the asymptotic behavior of spherical integrals for rectangular matrices and of low-rank matrix perturbations; the second one relies on the design and analysis of an auxiliary AMP. The numerical experiments show that there is a performance gap between the AMP and Bayes estimators, which is due to the incorrect estimation of the signal norm.

cs.IT↗

arXiv2022

Approximate Message Passing for Multi-Layer Estimation in Rotationally Invariant Models

We consider the problem of reconstructing the signal and the hidden variables from observations coming from a multi-layer network with rotationally invariant weight matrices. The multi-layer structure models inference from deep generative priors, and the rotational invariance imposed on the weights generalizes the i.i.d.\ Gaussian assumption by allowing for a complex correlation structure, which is typical in applications. In this work, we present a new class of approximate message passing (AMP) algorithms and give a state evolution recursion which precisely characterizes their performance in the large system limit. In contrast with the existing multi-layer VAMP (ML-VAMP) approach, our proposed AMP -- dubbed multi-layer rotationally invariant generalized AMP (ML-RI-GAMP) -- provides a natural generalization beyond Gaussian designs, in the sense that it recovers the existing Gaussian AMP as a special case. Furthermore, ML-RI-GAMP exhibits a significantly lower complexity than ML-VAMP, as the computationally intensive singular value decomposition is replaced by an estimation of the moments of the design matrices. Finally, our numerical results show that this complexity gain comes at little to no cost in the performance of the algorithm.

stat.ML↗

arXiv2022

Sparse superposition codes with rotational invariant coding matrices for memoryless channels

We recently showed in [1] the superiority of certain structured coding matrices ensembles (such as partial row-orthogonal) for sparse superposition codes when compared with purely random matrices with i.i.d. entries, both information-theoretically and under practical vector approximate message-passing decoding. Here we generalize this result to binary input channels under generalized vector approximate message-passing decoding [2].We focus on specific binary output channels for concreteness but our analysis based on the replica symmetric method from statistical physics applies to any memoryless channel. We confirm that the "spectral criterion" introduced in [1], a coding-matrix design principle which allows the code to be capacity-achieving in the "large section size" asymptotic limit, extends to generic memoryless channels. Moreover, we also show that the vanishing error floor property [3] of this coding scheme is universal for arbitrary spectrum of the coding matrix.

cs.IT↗

arXiv2022

Sparse superposition codes under VAMP decoding with generic rotational invariant coding matrices

Sparse superposition codes were originally proposed as a capacity-achieving communication scheme over the gaussian channel, whose coding matrices were made of i.i.d. gaussian entries.We extend this coding scheme to more generic ensembles of rotational invariant coding matrices with arbitrary spectrum, which include the gaussian ensemble as a special case. We further introduce and analyse a decoder based on vector approximate message-passing (VAMP).Our main findings, based on both a standard replica symmetric potential theory and state evolution analysis, are the superiority of certain structured ensembles of coding matrices (such as partial row-orthogonal) when compared to i.i.d. matrices, as well as a spectrum-independent upper bound on VAMP's threshold. Most importantly, we derive a simple "spectral criterion " for the scheme to be at the same time capacity-achieving while having the best possible algorithmic threshold, in the "large section size" asymptotic limit. Our results therefore provide practical design principles for the coding matrices in this promising communication scheme.

cs.IT↗

arXiv2022

The price of ignorance: how much does it cost to forget noise structure in low-rank matrix estimation?

We consider the problem of estimating a rank-1 signal corrupted by structured rotationally invariant noise, and address the following question: how well do inference algorithms perform when the noise statistics is unknown and hence Gaussian noise is assumed? While the matched Bayes-optimal setting with unstructured noise is well understood, the analysis of this mismatched problem is only at its premises. In this paper, we make a step towards understanding the effect of the strong source of mismatch which is the noise statistics. Our main technical contribution is the rigorous analysis of a Bayes estimator and of an approximate message passing (AMP) algorithm, both of which incorrectly assume a Gaussian setup. The first result exploits the theory of spherical integrals and of low-rank matrix perturbations; the idea behind the second one is to design and analyze an artificial AMP which, by taking advantage of the flexibility in the denoisers, is able to "correct" the mismatch. Armed with these sharp asymptotic characterizations, we unveil a rich and often unexpected phenomenology. For example, despite AMP is in principle designed to efficiently compute the Bayes estimator, the former is outperformed by the latter in terms of mean-square error. We show that this performance gap is due to an incorrect estimation of the signal norm. In fact, when the SNR is large enough, the overlaps of the AMP and the Bayes estimator coincide, and they even match those of optimal estimators taking into account the structure of the noise.

cs.IT↗