arXiv subjects

Feng Wu

Publications and source records attributed to Feng Wu.

At least 127 records · Page 7Linked to original sources

arXiv2024

Uniformly Accelerated Motion Model for Inter Prediction

Inter prediction is a key technology to reduce the temporal redundancy in video coding. In natural videos, there are usually multiple moving objects with variable velocity, resulting in complex motion fields that are difficult to represent compactly. In Versatile Video Coding (VVC), existing inter prediction methods usually assume uniform speed motion between consecutive frames and use the linear models for motion estimation (ME) and motion compensation (MC), which may not well handle the complex motion fields in the real world. To address these issues, we introduce a uniformly accelerated motion model (UAMM) to exploit motion-related elements (velocity, acceleration) of moving objects between the video frames, and further combine them to assist the inter prediction methods to handle the variable motion in the temporal domain. Specifically, first, the theory of UAMM is mentioned. Second, based on that, we propose the UAMM-based parameter derivation and extrapolation schemes in the coding process. Third, we integrate the UAMM into existing inter prediction modes (Merge, MMVD, CIIP) to achieve higher prediction accuracy. The proposed method is implemented into the VVC reference software, VTM version 12.0. Experimental results show that the proposed method achieves up to 0.38% and on average 0.13% BD-rate reduction compared to the VTM anchor, under the Low-delay P configuration, with a slight increase of time complexity on the encoding/decoding side.

eess.IV↗

arXiv2024

Inference Attacks: A Taxonomy, Survey, and Promising Directions

The prosperity of machine learning has also brought people's concerns about data privacy. Among them, inference attacks can implement privacy breaches in various MLaaS scenarios and model training/prediction phases. Specifically, inference attacks can perform privacy inference on undisclosed target training sets based on outputs of the target model, including but not limited to statistics, membership, semantics, data representation, etc. For instance, infer whether the target data has the characteristics of AIDS. In addition, the rapid development of the machine learning community in recent years, especially the surge of model types and application scenarios, has further stimulated the inference attacks' research. Thus, studying inference attacks and analyzing them in depth is urgent and significant. However, there is still a gap in the systematic discussion of inference attacks from taxonomy, global perspective, attack, and defense perspectives. This survey provides an in-depth and comprehensive inference of attacks and corresponding countermeasures in ML-as-a-service based on taxonomy and the latest researches. Without compromising researchers' intuition, we first propose the 3MP taxonomy based on the community research status, trying to normalize the confusing naming system of inference attacks. Also, we analyze the pros and cons of each type of inference attack, their workflow, countermeasure, and how they interact with other attacks. In the end, we point out several promising directions for researchers from a more comprehensive and novel perspective.

cs.LG↗

arXiv2024

Towards Harmonization of SO(3)-Equivariance and Expressiveness: a Hybrid Deep Learning Framework for Electronic-Structure Hamiltonian Prediction

Deep learning for predicting the electronic-structure Hamiltonian of quantum systems necessitates satisfying the covariance laws, among which achieving SO(3)-equivariance without sacrificing the non-linear expressive capability of networks remains unsolved. To navigate the harmonization between equivariance and expressiveness, we propose a deep learning method synergizing two distinct categories of neural mechanisms as a two-stage encoding and regression framework. The first stage corresponds to group theory-based neural mechanisms with inherent SO(3)-equivariant properties prior to the parameter learning process, while the second stage is characterized by a non-linear 3D graph Transformer network we propose, featuring high capability on non-linear expressiveness. The novel combination lies in the point that, the first stage predicts baseline Hamiltonians with abundant SO(3)-equivariant features extracted, assisting the second stage in empirical learning of equivariance; and in turn, the second stage refines the first stage's output as a fine-grained prediction of Hamiltonians using powerful non-linear neural mappings, compensating for the intrinsic weakness on non-linear expressiveness capability of mechanisms in the first stage. Our method enables precise, generalizable predictions while capturing SO(3)-equivariance under rotational transformations, and achieves state-of-the-art performance in Hamiltonian prediction on six benchmark databases.

physics.comp-ph↗

arXiv2024

Enhanced second harmonic generation in high-$Q$ all-dielectric metasurfaces with backward frequency conversion

Here we employ the quasi-bound state in the continuum (quasi-BIC) resonance in all-dielectric metasurfaces for efficient nonlinear processes in consideration of the backward frequency conversion. We theoretically study the second-harmonic generation (SHG) from symmetry-broken AlGaAs metasurfaces and reveal the efficiency enhancement empowered by high-$Q$ quasi-BIC resonances. By introducing the correction term of nonlinear polarization at the fundamental wave field to the conventional undepleted approximation, we uncover the effect of backward frequency conversion on the nonlinear conversation efficiency. The SHG efficiency as $2.45\times10^{-2}$ with the developed depleted model, shows a $14.3\%$ decrease compared with $2.86\times10^{-2}$ in conventional undepleted approximation, under the incident intensity of 10 MW/cm$^{2}$. Our results are of significant importance for designing efficient nonlinear metasurfaces supporting high-$Q$ resonances.

physics.optics↗

arXiv2024

ROPO: Robust Preference Optimization for Large Language Models

Preference alignment is pivotal for empowering large language models (LLMs) to generate helpful and harmless responses. However, the performance of preference alignment is highly sensitive to the prevalent noise in the preference data. Recent efforts for this problem either marginally alleviate the impact of noise without the ability to actually reduce its presence, or rely on costly teacher LLMs prone to reward misgeneralization. To address these challenges, we propose the RObust Preference Optimization (ROPO) framework, an iterative alignment approach that integrates noise-tolerance and filtering of noisy samples without the aid of external models. Specifically, ROPO iteratively solves a constrained optimization problem, where we dynamically assign a quality-aware weight for each sample and constrain the sum of the weights to the number of samples we intend to retain. For noise-tolerant training and effective noise identification, we derive a robust loss by suppressing the gradients of samples with high uncertainty. We demonstrate both empirically and theoretically that the derived loss is critical for distinguishing noisy samples from clean ones. Furthermore, inspired by our derived loss, we propose a robustness-guided rejection sampling technique to compensate for the potential important information in discarded queries. Experiments on three widely-used datasets with Mistral-7B and Llama-2-7B demonstrate that ROPO significantly outperforms existing preference alignment methods, with its superiority growing as the noise rate increases.

cs.LG↗

arXiv2024

Refractive index sensing based on large negative Goos-H$\ddot{a}$nchen shifts of wavy dielectric grating

Wavy dielectric grating hosts bound states in the continuum (BICs) at nonzero Bloch wave number. For oblique incident optical field with parameters near to the BICs, the reflectance spectrum exhibits ultra-sharp Fano line shape, and the reflected beam has large negative Goos-H$\ddot{a}$nchen shift, due to excitation of the corresponding quasi-BIC with negative group velocity. Under incidence of Gaussian beam with sizable beam width, the excited quasi-BIC could travel a long distance along the direction of the Goos-H$\ddot{a}$nchen shift, designated as $L_{GH}$, before the energy is completely radiated. If the length between the termination of the wavy shape and the focus of the incident Gaussian beam is smaller than $L_{GH}$, sizable energy flux can be coupled into the waveguide mode of the flat dielectric slab that is connected to the wavy dielectric grating. Measurement of the energy flux of the waveguide mode can sense the variation of the refractive index of the background medium. The proposed sensing scheme can be integrated with waveguide in optical circuit.

physics.optics↗

arXiv2024

Structured Click Control in Transformer-based Interactive Segmentation

Click-point-based interactive segmentation has received widespread attention due to its efficiency. However, it's hard for existing algorithms to obtain precise and robust responses after multiple clicks. In this case, the segmentation results tend to have little change or are even worse than before. To improve the robustness of the response, we propose a structured click intent model based on graph neural networks, which adaptively obtains graph nodes via the global similarity of user-clicked Transformer tokens. Then the graph nodes will be aggregated to obtain structured interaction features. Finally, the dual cross-attention will be used to inject structured interaction features into vision Transformer features, thereby enhancing the control of clicks over segmentation results. Extensive experiments demonstrated the proposed algorithm can serve as a general structure in improving Transformer-based interactive segmenta?tion performance. The code and data will be released at https://github.com/hahamyt/scc.

cs.CV↗

arXiv2024

DMOFC: Discrimination Metric-Optimized Feature Compression

Feature compression, as an important branch of video coding for machines (VCM), has attracted significant attention and exploration. However, the existing methods mainly focus on intra-feature similarity, such as the Mean Squared Error (MSE) between the reconstructed and original features, while neglecting the importance of inter-feature relationships. In this paper, we analyze the inter-feature relationships, focusing on feature discriminability in machine vision and underscoring its significance in feature compression. To maintain the feature discriminability of reconstructed features, we introduce a discrimination metric for feature compression. The discrimination metric is designed to ensure that the distance between features of the same category is smaller than the distance between features of different categories. Furthermore, we explore the relationship between the discrimination metric and the discriminability of the original features. Experimental results confirm the effectiveness of the proposed discrimination metric and reveal there exists a trade-off between the discrimination metric and the discriminability of the original features.

cs.CV↗

arXiv2024

Improving Offline Reinforcement Learning with Inaccurate Simulators

Offline reinforcement learning (RL) provides a promising approach to avoid costly online interaction with the real environment. However, the performance of offline RL highly depends on the quality of the datasets, which may cause extrapolation error in the learning process. In many robotic applications, an inaccurate simulator is often available. However, the data directly collected from the inaccurate simulator cannot be directly used in offline RL due to the well-known exploration-exploitation dilemma and the dynamic gap between inaccurate simulation and the real environment. To address these issues, we propose a novel approach to combine the offline dataset and the inaccurate simulation data in a better manner. Specifically, we pre-train a generative adversarial network (GAN) model to fit the state distribution of the offline dataset. Given this, we collect data from the inaccurate simulator starting from the distribution provided by the generator and reweight the simulated data using the discriminator. Our experimental results in the D4RL benchmark and a real-world manipulation task confirm that our method can benefit more from both inaccurate simulator and limited offline datasets to achieve better performance than the state-of-the-art methods.

cs.RO↗

arXiv2024

Learning to Cut via Hierarchical Sequence/Set Model for Efficient Mixed-Integer Programming

Cutting planes (cuts) play an important role in solving mixed-integer linear programs (MILPs), which formulate many important real-world applications. Cut selection heavily depends on (P1) which cuts to prefer and (P2) how many cuts to select. Although modern MILP solvers tackle (P1)-(P2) by human-designed heuristics, machine learning carries the potential to learn more effective heuristics. However, many existing learning-based methods learn which cuts to prefer, neglecting the importance of learning how many cuts to select. Moreover, we observe that (P3) what order of selected cuts to prefer significantly impacts the efficiency of MILP solvers as well. To address these challenges, we propose a novel hierarchical sequence/set model (HEM) to learn cut selection policies. Specifically, HEM is a bi-level model: (1) a higher-level module that learns how many cuts to select, (2) and a lower-level module -- that formulates the cut selection as a sequence/set to sequence learning problem -- to learn policies selecting an ordered subset with the cardinality determined by the higher-level module. To the best of our knowledge, HEM is the first data-driven methodology that well tackles (P1)-(P3) simultaneously. Experiments demonstrate that HEM significantly improves the efficiency of solving MILPs on eleven challenging MILP benchmarks, including two Huawei's real problems.

cs.AI↗

arXiv2024

Autothrottle: A Practical Bi-Level Approach to Resource Management for SLO-Targeted Microservices

Achieving resource efficiency while preserving end-user experience is non-trivial for cloud application operators. As cloud applications progressively adopt microservices, resource managers are faced with two distinct levels of system behavior: end-to-end application latency and per-service resource usage. Translating between the two levels, however, is challenging because user requests traverse heterogeneous services that collectively (but unevenly) contribute to the end-to-end latency. We present Autothrottle, a bi-level resource management framework for microservices with latency SLOs (service-level objectives). It architecturally decouples application SLO feedback from service resource control, and bridges them through the notion of performance targets. Specifically, an application-wide learning-based controller is employed to periodically set performance targets -- expressed as CPU throttle ratios -- for per-service heuristic controllers to attain. We evaluate Autothrottle on three microservice applications, with workload traces from production scenarios. Results show superior CPU savings, up to 26.21% over the best-performing baseline and up to 93.84% over all baselines.

cs.DC↗

arXiv2024

Scene Adaptive Sparse Transformer for Event-based Object Detection

While recent Transformer-based approaches have shown impressive performances on event-based object detection tasks, their high computational costs still diminish the low power consumption advantage of event cameras. Image-based works attempt to reduce these costs by introducing sparse Transformers. However, they display inadequate sparsity and adaptability when applied to event-based object detection, since these approaches cannot balance the fine granularity of token-level sparsification and the efficiency of window-based Transformers, leading to reduced performance and efficiency. Furthermore, they lack scene-specific sparsity optimization, resulting in information loss and a lower recall rate. To overcome these limitations, we propose the Scene Adaptive Sparse Transformer (SAST). SAST enables window-token co-sparsification, significantly enhancing fault tolerance and reducing computational overhead. Leveraging the innovative scoring and selection modules, along with the Masked Sparse Window Self-Attention, SAST showcases remarkable scene-aware adaptability: It focuses only on important objects and dynamically optimizes sparsity level according to scene complexity, maintaining a remarkable balance between performance and computational cost. The evaluation results show that SAST outperforms all other dense and sparse networks in both performance and efficiency on two large-scale event-based object detection datasets (1Mpx and Gen1). Code: https://github.com/Peterande/SAST

cs.CV↗

arXiv2024

The Four-Boson First-Excited State Near Two-Body Unitarity

Near two-body unitarity, the three-boson system is characterized by an approximate discrete scale invariance manifest in a geometric tower of bound states (the Efimov effect). In the absence of a strong four-body force, the four-boson system has two states associated with each Efimov state, one very nearly unstable, the other several times deeper. We study correlations between the excited- and ground-state properties, such as binding energies and radii, up to next-to-leading order in an effective field theory for short-range forces. We obtain the parameters in these correlations from similar correlations arising from existing precise calculations based on short-range potentials. We also derive correlations among excited-state properties that emerge from the proximity of the state to the break-up threshold into a boson and a three-boson bound state, using an effective field theory for "halo" states.

physics.atom-ph↗

arXiv2024

Accelerating Data Generation for Neural Operators via Krylov Subspace Recycling

Learning neural operators for solving partial differential equations (PDEs) has attracted great attention due to its high inference efficiency. However, training such operators requires generating a substantial amount of labeled data, i.e., PDE problems together with their solutions. The data generation process is exceptionally time-consuming, as it involves solving numerous systems of linear equations to obtain numerical solutions to the PDEs. Many existing methods solve these systems independently without considering their inherent similarities, resulting in extremely redundant computations. To tackle this problem, we propose a novel method, namely Sorting Krylov Recycling (SKR), to boost the efficiency of solving these systems, thus significantly accelerating data generation for neural operators training. To the best of our knowledge, SKR is the first attempt to address the time-consuming nature of data generation for learning neural operators. The working horse of SKR is Krylov subspace recycling, a powerful technique for solving a series of interrelated systems by leveraging their inherent similarities. Specifically, SKR employs a sorting algorithm to arrange these systems in a sequence, where adjacent systems exhibit high similarities. Then it equips a solver with Krylov subspace recycling to solve the systems sequentially instead of independently, thus effectively enhancing the solving efficiency. Both theoretical analysis and extensive experiments demonstrate that SKR can significantly accelerate neural operator data generation, achieving a remarkable speedup of up to 13.9 times.

cs.LG↗

arXiv2024

A Deep Instance Generative Framework for MILP Solvers Under Limited Data Availability

In the past few years, there has been an explosive surge in the use of machine learning (ML) techniques to address combinatorial optimization (CO) problems, especially mixed-integer linear programs (MILPs). Despite the achievements, the limited availability of real-world instances often leads to sub-optimal decisions and biased solver assessments, which motivates a suite of synthetic MILP instance generation techniques. However, existing methods either rely heavily on expert-designed formulations or struggle to capture the rich features of real-world instances. To tackle this problem, we propose G2MILP, the first deep generative framework for MILP instances. Specifically, G2MILP represents MILP instances as bipartite graphs, and applies a masked variational autoencoder to iteratively corrupt and replace parts of the original graphs to generate new ones. The appealing feature of G2MILP is that it can learn to generate novel and realistic MILP instances without prior expert-designed formulations, while preserving the structures and computational hardness of real-world datasets, simultaneously. Thus the generated instances can facilitate downstream tasks for enhancing MILP solvers under limited data availability. We design a suite of benchmarks to evaluate the quality of the generated MILP instances. Experiments demonstrate that our method can produce instances that closely resemble real-world datasets in terms of both structures and computational hardness. The deliverables are released at https://miralab-ustc.github.io/L2O-G2MILP.

cs.LG↗

arXiv2024

Efficient initialization of fluxonium qubits based on auxiliary energy levels

Fast and high-fidelity qubit initialization is crucial for low-frequency qubits such as fluxonium, and in applications of many quantum algorithms and quantum error correction codes. In a circuit quantum electrodynamics system, the initialization is typically achieved by transferring the state between the qubit and a short-lived cavity through microwave driving, also known as the sideband cooling process in atomic system. Constrained by the selection rules from the parity symmetry of the wavefunctions, the sideband transitions are only enabled by multi-photon processes which requires multi-tone or strong driving. Leveraging the flux-tunability of fluxonium, we circumvent this limitation by breaking flux symmetry to enable an interaction between a non-computational qubit transition and the cavity excitation. With single-tone sideband driving, we realize qubit initialization with a fidelity exceeding 99% within a duration of 300 ns, robust against the variation of control parameters. Furthermore, we show that our initialization scheme has a built-in benefit in simultaneously removing the second-excited state population of the qubit, and can be easily incorporated into a large-scale fluxonium processor.

quant-ph↗

arXiv2024

Efficient photon-pair generation empowered by dual quasi-bound states in the continuum

Here we demonstrate the efficient photon-pair generation via spontaneous parametric down conversion from a semiconductor metasurface supporting dual quasi-bound states in the continuum (quasi-BICs). In a simple metasurface design composed of AlGaAs ellipse nano-cyclinders, the two high-$Q$ quasi-BIC resonances that coincide with the generated signal and idler frequencies significantly boost the local electric field. This leads to a substantial enhancement in the reverse classical nonlinear process of sum frequency generation and subsequently the remarkable high generation rate of photon pairs under the quantum-classical correspondence principle. Within a narrowband wavelength regime around the quasi-BIC resonances, the rate of pair production is enhanced up to $\sim10^{4}$ Hz, two orders of magnitude larger than that in the Mie resonant AlGaAs nanoantennas. Moreover, the photon pair emission is mainly concentrated in the normal direction with respect to the metasurface, and shows tunable rate with the $Q$ factor by engineering the rotation angle of nano-cylinders. The presented work enables nanoscale sources of high-quality entangled photons which will find applications in advanced quantum imaging and communications.

physics.optics↗

arXiv2024

Graph Relation Distillation for Efficient Biomedical Instance Segmentation

Instance-aware embeddings predicted by deep neural networks have revolutionized biomedical instance segmentation, but its resource requirements are substantial. Knowledge distillation offers a solution by transferring distilled knowledge from heavy teacher networks to lightweight yet high-performance student networks. However, existing knowledge distillation methods struggle to extract knowledge for distinguishing instances and overlook global relation information. To address these challenges, we propose a graph relation distillation approach for efficient biomedical instance segmentation, which considers three essential types of knowledge: instance-level features, instance relations, and pixel-level boundaries. We introduce two graph distillation schemes deployed at both the intra-image level and the inter-image level: instance graph distillation (IGD) and affinity graph distillation (AGD). IGD constructs a graph representing instance features and relations, transferring these two types of knowledge by enforcing instance graph consistency. AGD constructs an affinity graph representing pixel relations to capture structured knowledge of instance boundaries, transferring boundary-related knowledge by ensuring pixel affinity consistency. Experimental results on a number of biomedical datasets validate the effectiveness of our approach, enabling student models with less than $ 1\%$ parameters and less than $10\%$ inference time while achieving promising performance compared to teacher models.

cs.CV↗