arXiv subjects

An Chen

Publications and source records attributed to An Chen.

At least 19 recordsLinked to original sources

arXiv2026

Rectify Then Diffuse: Disentangling Concepts Before Denoising Trajectory Unfolds

Text-to-image diffusion models can generate individual concepts well, but they often omit or merge concepts incorrectly with multiple concepts. We trace these failures to an early coordination bottleneck: before denoising begins, prompt-conditioned attention may allocate different concepts to strongly overlapping spatial support, which can keep their attention coupled as denoising proceeds. This observation motivates treating compositional generation as a boundary-condition problem rather than repeatedly controlling the evolving trajectory. To this end, we propose Rectify-then-Diffuse (RTD), a training-free framework that rectifies the initial allocation once before standard denoising. Firstly, we propose Soft-Overlap Disentanglement (SOD), which converts normalized overlap between pilot concept maps into a differentiable and layout-agnostic separation objective. Secondly, we introduce Isotropic Gradient Rectification (IGR), which normalizes the SOD gradient and applies a bounded latent displacement with a consistent scale across prompts and initializations. Extensive experiments show that RTD achieves state-of-the-art compositional fidelity and robust gains. On the AE-Bench object pair subset, RTD improves BLIP-VQA by 45.8% and ImageReward by 19.6% over CO3 while running 2.3$\times$ faster. Code will be released at https://github.com/Z-yiwei/rectify-then-diffuse

cs.CV↗

arXiv2026

One Knob to Rule Them All: A Unified Optimal Transport View of Cold-Start Active Learning

Cold-Start Active Learning (CSAL) aims to select a valuable subset from an unlabeled pool without any prior knowledge or human assistance. Existing methods take diverse routes based on typicality, coverage, or diversity. Each rests on its own inductive bias and therefore performs well on some tasks yet poorly on others. We argue that the real challenge is not to design yet another selection heuristic, but to make CSAL adapt automatically to the data and task at hand. To this end, we revisit CSAL through the lens of optimal transport. First, we propose a generalized transport selection framework that reveals the shared allocation structure of existing methods and exactly subsumes representative formulations. Second, we introduce a theoretical analysis that characterizes the trade-off controlled by entropic regularization and establishes a task-agnostic minimax bound for cold-start selection. These results provide a principled foundation for adapting the regularization strength to the unlabeled data. Third, we derive a data-adaptive regularization rule and present a novel Sinkhorn-based CSAL algorithm, termed $\epsilon$-Adaptive Selection ($\epsilon$-AS). Extensive experiments on six public datasets and multiple annotation budgets show that $\epsilon$-AS consistently achieves state-of-the-art performance. On ImageNet-1k, it improves the average accuracy over ActiveFT by 1.29% while reducing selection time by 56.2%. Code will be released at https://github.com/Z-yiwei/OT-CSAL

cs.AI↗

arXiv2026

Sufficient Conditions for Unique Optimizer of Two-Dimensional Atomic Norm Minimization Under Multiple Frequencies

Atomic norm minimization (ANM) has been extensively applied for gridless angle estimation. However, with the increase of the number of antennas and the communication frequencies in massive MIMO systems, the accompanying beam squint effect significantly degrades angle estimation accuracy. Existing solutions either address this issue only in the one-dimensional (1D) SIMO case, or decouple the two-dimensional (2D) angle estimation into two separate 1D problems, which fails to achieve the optimal solution. In this paper, we employ the multi-frequency model to characterize the beam squint effect in MIMO channels and propose a multi-frequency version of the ANM objective for corresponding 2D angle estimation. To efficiently retrieve the angle parameters, we prove the existence of the equivalent semi-definite program formulation of the ANM objective and develop an algorithm based on the alternating direction method of multipliers for its solutions. Moreover, we derive the certification conditions of this objective to guarantee the existence of a unique optimal solution.

eess.SP↗

arXiv2026

Nonconcave Portfolio Choice under Smooth Ambiguity

We study continuous-time portfolio choice with nonlinear payoffs under smooth ambiguity and Bayesian learning. We develop a general framework for dynamic, non-concave asset allocation that accommodates nonlinear payoffs, broad utility classes, and flexible ambiguity attitudes. Dynamic consistency is obtained by a robust representation that recasts the ambiguity-averse problem as ambiguity-neutral with distorted priors. This structure delivers explicit trading rules by combining nonlinear filtering with the martingale approach and nests standard concave and linear-payoff benchmarks. As a leading application, delegated management with convex incentives illustrates that ambiguity aversion shifts beliefs toward adverse states, limits the range of states that would otherwise trigger more aggressive risk taking, and reduces volatility through lower risky exposure.

q-fin.PM↗

arXiv2025

Unifying Block-wise PTQ and Distillation-based QAT for Progressive Quantization toward 2-bit Instruction-Tuned LLMs

As the rapid scaling of large language models (LLMs) poses significant challenges for deployment on resource-constrained devices, there is growing interest in extremely low-bit quantization, such as 2-bit. Although prior works have shown that 2-bit large models are pareto-optimal over their 4-bit smaller counterparts in both accuracy and latency, these advancements have been limited to pre-trained LLMs and have not yet been extended to instruction-tuned models. To bridge this gap, we propose Unified Progressive Quantization (UPQ)$-$a novel progressive quantization framework (FP16$\rightarrow$INT4$\rightarrow$INT2) that unifies block-wise post-training quantization (PTQ) with distillation-based quantization-aware training (Distill-QAT) for INT2 instruction-tuned LLM quantization. UPQ first quantizes FP16 instruction-tuned models to INT4 using block-wise PTQ to significantly reduce the quantization error introduced by subsequent INT2 quantization. Next, UPQ applies Distill-QAT to enable INT2 instruction-tuned LLMs to generate responses consistent with their original FP16 counterparts by minimizing the generalized Jensen-Shannon divergence (JSD) between the two. To the best of our knowledge, we are the first to demonstrate that UPQ can quantize open-source instruction-tuned LLMs to INT2 without relying on proprietary post-training data, while achieving state-of-the-art performances on MMLU and IFEval$-$two of the most representative benchmarks for evaluating instruction-tuned LLMs.

cs.LG↗

arXiv2025

Observation of Coherent Perfect Acoustic Absorption at an Exceptional Point

Non-Hermitian systems have recently shown new possibilities to manipulate wave scattering by exploiting loss, yet coherent perfect absorption at an exceptional point (CPA EP) remains elusive in acoustics. Here we demonstrate it based on a two-channel waveguide with compact lossy resonators. We realize imbalanced losses crucial for CPA EP by using active components to independently modulate the non-Hermiticity. The CPA EP experimentally manifests as full absorption at a unique real frequency and shows high sensitivity to the incident phase variations.Our findings open an avenue to explore novel non-Hermitian physics for classical waves and develop innovative acoustic singularity-based devices.

physics.app-ph↗

arXiv2025

Analog Foundation Models

Analog in-memory computing (AIMC) is a promising compute paradigm to improve speed and power efficiency of neural network inference beyond the limits of conventional von Neumann-based architectures. However, AIMC introduces fundamental challenges such as noisy computations and strict constraints on input and output quantization. Because of these constraints and imprecisions, off-the-shelf LLMs are not able to achieve 4-bit-level performance when deployed on AIMC-based hardware. While researchers previously investigated recovering this accuracy gap on small, mostly vision-based models, a generic method applicable to LLMs pre-trained on trillions of tokens does not yet exist. In this work, we introduce a general and scalable method to robustly adapt LLMs for execution on noisy, low-precision analog hardware. Our approach enables state-of-the-art models $\unicode{x2013}$ including Phi-3-mini-4k-instruct and Llama-3.2-1B-Instruct $\unicode{x2013}$ to retain performance comparable to 4-bit weight, 8-bit activation baselines, despite the presence of analog noise and quantization constraints. Additionally, we show that as a byproduct of our training methodology, analog foundation models can be quantized for inference on low-precision digital hardware. Finally, we show that our models also benefit from test-time compute scaling, showing better scaling behavior than models trained with 4-bit weight and 8-bit static input quantization. Our work bridges the gap between high-capacity LLMs and efficient analog hardware, offering a path toward energy-efficient foundation models. Code is available at https://github.com/IBM/analog-foundation-models.

cs.LG↗

arXiv2025

Acoustic topological Jackiw-Rebbi states at symmetry broken interfaces

Topological insulators, a fundamental concept in modern condensed matter physics, support localized states at the interfaces between insulators exhibiting different topological phases, which conventionally rely on explicit symmetry breaking. Here, we propose a mechanism to induce a real-space topological phase transition by spontaneous symmetry breaking, thereby constructing an acoustic metagrating to generate nontrivial Jackiw-Rebbi states characterized by robust imaginary band degeneracy. Our experimental implementation verifies the acoustic delocalized interface state with a constant phase jump, demonstrating enhanced high-directivity topological radiation. We foresee that our findings may spark interest in engineering new acoustic topological devices.

cond-mat.mes-hall↗

arXiv2025

FedX: Adaptive Model Decomposition and Quantization for IoT Federated Learning

Federated Learning (FL) allows collaborative training among multiple devices without data sharing, thus enabling privacy-sensitive applications on mobile or Internet of Things (IoT) devices, such as mobile health and asset tracking. However, designing an FL system with good model utility that works with low computation/communication overhead on heterogeneous, resource-constrained mobile/IoT devices is challenging. To address this problem, this paper proposes FedX, a novel adaptive model decomposition and quantization FL system for IoT. To balance utility with resource constraints on IoT devices, FedX decomposes a global FL model into different sub-networks with adaptive numbers of quantized bits for different devices. The key idea is that a device with fewer resources receives a smaller sub-network for lower overhead but utilizes a larger number of quantized bits for higher model utility, and vice versa. The quantization operations in FedX are done at the server to reduce the computational load on devices. FedX iteratively minimizes the losses in the devices' local data and in the server's public data using quantized sub-networks under a regularization term, and thus it maximizes the benefits of combining FL with model quantization through knowledge sharing among the server and devices in a cost-effective training process. Extensive experiments show that FedX significantly improves quantization times by up to 8.43X, on-device computation time by 1.5X, and total end-to-end training time by 1.36X, compared with baseline FL systems. We guarantee the global model convergence theoretically and validate local model convergence empirically, highlighting FedX's optimization efficiency.

cs.LG↗

arXiv2025

Molecular HDD Logic for Encrypted Massive Data Storage

Organic memories, with small dimension, fast speed and long retention features, are considered as promising candidates for massive data archiving. In order to satisfy the re-quirements for ultra-low power and high-security information storage, we design a concep-tual molecular hard-disk (HDD) logic scheme that is capable to execute in-situ encryption of massive data in pW/bit power-consumption range. Beneficial from the coupled mechanism of counter-balanced redox reaction and local ion drifting, the basic HDD unit consisting of ~ 200 self-assembled RuXLPH molecules in a monolayer (SAM) configuration undergoes unique conductance modulation with continuous, symmetric and low-power switching char-acteristics. 96-state memory performance, which allows 6-bit data storage and single-unit one-step XOR operation, is realized in the RuXLPH SAM sample. Through single-unit XOR manipulation of the pixel information, in-situ bitwise encryption of the Mogao Grottoes mural images stored in the molecular HDD is demonstrated.

cs.ET↗

arXiv2024

Birkhoff sum convergence of Fr\'echet observables to stable laws for Gibbs-Markov systems and applications

We use a Poisson point process approach to prove distributional convergence to a stable law for non square-integrable observables $\phi: [0,1]\to R$, mostly of the form $\phi (x) = d(x,x_0)^{-\frac{1}{\alpha}}$,$0<\alpha\le 2$, on Gibbs-Markov maps. A key result is to verify a standard mixing condition, which ensures that large values of the observable dominate the time-series, in the range $1<\alpha \le 2$. Stable limit laws for observables on dynamical systems have been established in two settings: ``good observables'' (typically H\"older) on slowly mixing non-uniformly hyperbolic systems and ``bad'' observables (unbounded with fat tails) on fast mixing dynamical systems. As an application we investigate the interplay between these two effects in a class of intermittent-type maps.

math.DS↗

arXiv2024

Knowledge-Reuse Transfer Learning Methods in Molecular and Material Science

Molecules and materials are the foundation for the development of modern advanced industries such as energy storage systems and semiconductor devices. However, traditional trial-and-error methods or theoretical calculations are highly resource-intensive, and extremely long R&D (Research and Development) periods cannot meet the urgent need for molecules/materials in industrial development. Machine learning (ML) methods based on big data are expected to break this dilemma. However, the difficulty in constructing large-scale datasets of new molecules/materials due to the high cost of data acquisition and annotation limits the development of machine learning. The application of transfer learning lowers the data requirements for model training, which makes transfer learning stand out in researches addressing data quality issues. In this review, we summarize recent advances in transfer learning related to molecular and materials science. We focus on the application of transfer learning methods for the discovery of advanced molecules/materials, particularly, the construction of transfer learning frameworks for different systems, and how transfer learning can enhance the performance of models. In addition, the challenges of transfer learning are also discussed.

cond-mat.mtrl-sci↗

arXiv2023

Striking the Balance: Life Insurance Timing and Asset Allocation in Financial Planning

This paper investigates the consumption and investment decisions of an individual facing uncertain lifespan and stochastic labor income within a Black-Scholes market framework. A key aspect of our study involves the agent's option to choose when to acquire life insurance for bequest purposes. We examine two scenarios: one with a fixed bequest amount and another with a controlled bequest amount. Applying duality theory and addressing free-boundary problems, we analytically solve both cases, and provide explicit expressions for value functions and optimal strategies in both cases. In the first scenario, where the bequest amount is fixed, distinct outcomes emerge based on different levels of risk aversion parameter $\gamma$: (i) the optimal time for life insurance purchase occurs when the agent's wealth surpasses a critical threshold if $\gamma \in (0,1)$, or (ii) life insurance should be acquired immediately if $\gamma>1$. In contrast, in the second scenario with a controlled bequest amount, regardless of $\gamma$ values, immediate life insurance purchase proves to be optimal. Finally, we extend the analysis to consider a scenario in which the individual earmarks part of her initial wealth for inheritance, where a critical wealth threshold consistently emerges.

q-fin.PM↗

arXiv2023

Sparsity-Based Channel Estimation Exploiting Deep Unrolling for Downlink Massive MIMO

Massive multiple-input multiple-output (MIMO) enjoys great advantage in 5G wireless communication systems owing to its spectrum and energy efficiency. However, hundreds of antennas require large volumes of pilot overhead to guarantee reliable channel estimation in FDD massive MIMO system. Compressive sensing (CS) has been applied for channel estimation by exploiting the inherent sparse structure of massive MIMO channel but suffer from high complexity. To overcome this challenge, this paper develops a hybrid channel estimation scheme by integrating the model-driven CS and data-driven deep unrolling technique. The proposed scheme consists of a coarse estimation part and a fine correction part to respectively exploit the inter- and intraframe sparsities of channels to greatly reduce the pilot overhead. Theoretical result is provided to indicate the convergence of the fine correction and coarse estimation net. Simulation results are provided to verify that our scheme can estimate MIMO channels with low pilot overhead while guaranteeing estimation accuracy with relatively low complexity.

eess.SP↗

arXiv2023

AlphaMat: A Material Informatics Hub Connecting Data, Features, Models and Applications

The development of modern civil industry, energy and information technology is inseparable from the rapid explorations of new materials, which are hampered by months to years of painstaking attempts, resulting in only a small fraction of materials being determined in a vast chemical space. Artificial intelligence (AI)-based methods are promising to address this gap, but face many challenges such as data scarcity and inaccurate material descriptor coding. Here, we develop an AI platform, AlphaMat, that connects materials and applications. AlphaMat is not limited by the data scale (from 101 to 106) and can design structural and component descriptors that are effective for docking with various AI models. With prediction time of milliseconds and high accuracy, AlphaMat exhibits strong powers to model at least 12 common attributes (formation energy, band gap, ionic conductivity, magnetism, phonon property, bulk modulus, dielectric constant, adsorption energy, etc.), resulting in an unexplored material database with over 117,000 entries. We further demonstrate the ability of AlphaMat to mine and design materials, which successfully discover thousands of new materials in photonics, batteries, catalysts, and capacitors from the largest inorganic compound databases that cover all elements in periodic table. This work proposes the first material informatics hub that does not require users to have strong programming knowledge to build AI models to design materials. Users can either directly retrieve our database or easily build AI models through AlphaMat to discover and design the required materials. AlphaMat can shorten the cycle of database construction and material discovery by at least decades, and its effective use will facilitate the applications of AI technology in material science and lead scientific and technological progress to a new height.

physics.chem-ph↗

arXiv2023

Zone-based Federated Learning for Mobile Sensing Data

Mobile apps, such as mHealth and wellness applications, can benefit from deep learning (DL) models trained with mobile sensing data collected by smart phones or wearable devices. However, currently there is no mobile sensing DL system that simultaneously achieves good model accuracy while adapting to user mobility behavior, scales well as the number of users increases, and protects user data privacy. We propose Zone-based Federated Learning (ZoneFL) to address these requirements. ZoneFL divides the physical space into geographical zones mapped to a mobile-edge-cloud system architecture for good model accuracy and scalability. Each zone has a federated training model, called a zone model, which adapts well to data and behaviors of users in that zone. Benefiting from the FL design, the user data privacy is protected during the ZoneFL training. We propose two novel zone-based federated training algorithms to optimize zone models to user mobility behavior: Zone Merge and Split (ZMS) and Zone Gradient Diffusion (ZGD). ZMS optimizes zone models by adapting the zone geographical partitions through merging of neighboring zones or splitting of large zones into smaller ones. Different from ZMS, ZGD maintains fixed zones and optimizes a zone model by incorporating the gradients derived from neighboring zones' data. ZGD uses a self-attention mechanism to dynamically control the impact of one zone on its neighbors. Extensive analysis and experimental results demonstrate that ZoneFL significantly outperforms traditional FL in two models for heart rate prediction and human activity recognition. In addition, we developed a ZoneFL system using Android phones and AWS cloud. The system was used in a heart rate prediction field study with 63 users for 4 months, and we demonstrated the feasibility of ZoneFL in real-life.

cs.LG↗

arXiv2023

A Surrogate-Assisted Highly Cooperative Coevolutionary Algorithm for Hyperparameter Optimization in Deep Convolutional Neural Network

Convolutional neural networks (CNNs) have gained remarkable success in recent years. However, their performance highly relies on the architecture hyperparameters, and finding proper hyperparameters for a deep CNN is a challenging optimization problem owing to its high-dimensional and computationally expensive characteristics. Given these difficulties, this study proposes a surrogate-assisted highly cooperative hyperparameter optimization (SHCHO) algorithm for chain-styled CNNs. To narrow the large search space, SHCHO first decomposes the whole CNN into several overlapping sub-CNNs in accordance with the overlapping hyperparameter interaction structure and then cooperatively optimizes these hyperparameter subsets. Two cooperation mechanisms are designed during this process. One coordinates all the sub-CNNs to reproduce the information flow in the whole CNN and achieve macro cooperation among them, and the other tackles the overlapping components by simultaneously considering the involved two sub-CNNs and facilitates micro cooperation between them. As a result, a proper hyperparameter configuration can be effectively located for the whole CNN. Besides, SHCHO also employs the well-performing surrogate technique to assist in the hyperparameter optimization of each sub-CNN, thereby greatly reducing the expensive computational cost. Extensive experimental results on two widely-used image classification datasets indicate that SHCHO can significantly improve the performance of CNNs.

cs.NE↗

arXiv2023

Hardware-aware training for large-scale and diverse deep learning inference workloads using in-memory computing-based accelerators

Analog in-memory computing (AIMC) -- a promising approach for energy-efficient acceleration of deep learning workloads -- computes matrix-vector multiplications (MVMs) but only approximately, due to nonidealities that often are non-deterministic or nonlinear. This can adversely impact the achievable deep neural network (DNN) inference accuracy as compared to a conventional floating point (FP) implementation. While retraining has previously been suggested to improve robustness, prior work has explored only a few DNN topologies, using disparate and overly simplified AIMC hardware models. Here, we use hardware-aware (HWA) training to systematically examine the accuracy of AIMC for multiple common artificial intelligence (AI) workloads across multiple DNN topologies, and investigate sensitivity and robustness to a broad set of nonidealities. By introducing a new and highly realistic AIMC crossbar-model, we improve significantly on earlier retraining approaches. We show that many large-scale DNNs of various topologies, including convolutional neural networks (CNNs), recurrent neural networks (RNNs), and transformers, can in fact be successfully retrained to show iso-accuracy on AIMC. Our results further suggest that AIMC nonidealities that add noise to the inputs or outputs, not the weights, have the largest impact on DNN accuracy, and that RNNs are particularly robust to all nonidealities.

cs.LG↗