Search

An ellipsometer is a vital precision tool used for measuring optical parameters with wide applications in many fields, including accurate measurements in film thickness, optical constants, structural profiles, etc. However, the precise measurement of photosensitive materials meets huge obstacles because of the excessive input photons, therefore the requirement of enhancing detection accuracy under low incident light intensity is an essential topic in the precision measurement. In this work, by combining a polarization-entangled photon source with a classical transmission-type ellipsometer, the quantum ellipsometer with the PSA (Polarizer-Sample-Analyzer) and the Senarmount method is constructed firstly to measure the phase retardation of the birefringent materials. The experimental results show that the accuracy can reach to nanometer scale at extremely low input intensity, and the stability are within 1% for all specimens tested with a compensator involved. Our work paves the way for precision measurement at low incident light intensity, with potential applications in measuring photosensitive materials, active-biological samples and other remote monitoring scenarios., Comment: 13 pages, 5 figures. This work has been submitted for possible publication

Foundation models have emerged as a powerful tool for many AI problems. Despite the tremendous success of foundation models, effective adaptation to new tasks, particularly those with limited labels, remains an open question and lacks theoretical understanding. An emerging solution with recent success in vision and NLP involves finetuning a foundation model on a selection of relevant tasks, before its adaptation to a target task with limited labeled samples. In this paper, we study the theoretical justification of this multitask finetuning approach. Our theoretical analysis reveals that with a diverse set of related tasks, this multitask finetuning leads to reduced error in the target task, in comparison to directly adapting the same pretrained model. We quantify the relationship between finetuning tasks and target tasks by diversity and consistency metrics, and further propose a practical task selection algorithm. We substantiate our theoretical claims with extensive empirical evidence. Further, we present results affirming our task selection algorithm adeptly chooses related finetuning tasks, providing advantages to the model performance on target tasks. We believe our study shed new light on the effective adaptation of foundation models to new tasks that lack abundant labels. Our code is available at https://github.com/OliverXUZY/Foudation-Model_Multitask., Comment: Published at ICLR 2024. 54 pages

Single object tracking (SOT) is a fundamental problem in computer vision, with a wide range of applications, including autonomous driving, augmented reality, and robot navigation. The robustness of SOT faces two main challenges: tiny target and fast motion. These challenges are especially manifested in videos captured by unmanned aerial vehicles (UAV), where the target is usually far away from the camera and often with significant motion relative to the camera. To evaluate the robustness of SOT methods, we propose BioDrone -- the first bionic drone-based visual benchmark for SOT. Unlike existing UAV datasets, BioDrone features videos captured from a flapping-wing UAV system with a major camera shake due to its aerodynamics. BioDrone hence highlights the tracking of tiny targets with drastic changes between consecutive frames, providing a new robust vision benchmark for SOT. To date, BioDrone offers the largest UAV-based SOT benchmark with high-quality fine-grained manual annotations and automatically generates frame-level labels, designed for robust vision analyses. Leveraging our proposed BioDrone, we conduct a systematic evaluation of existing SOT methods, comparing the performance of 20 representative models and studying novel means of optimizing a SOTA method (KeepTrack KeepTrack) for robust SOT. Our evaluation leads to new baselines and insights for robust SOT. Moving forward, we hope that BioDrone will not only serve as a high-quality benchmark for robust SOT, but also invite future research into robust computer vision. The database, toolkits, evaluation server, and baseline results are available at http://biodrone.aitestunion.com., Comment: This paper is published in IJCV (refer to DOI). Please cite the published IJCV

Objective: Subcutaneous Immunotherapy (SCIT) is the long-lasting causal treatment of allergic rhinitis (AR). How to enhance the adherence of patients to maximize the benefit of allergen immunotherapy (AIT) plays a crucial role in the management of AIT. This study aims to leverage novel machine learning models to precisely predict the risk of non-adherence of AR patients and related local symptom scores in three years SCIT. Methods: The research develops and analyzes two models, sequential latent-variable model (SLVM) of Stochastic Latent Actor-Critic (SLAC) and Long Short-Term Memory (LSTM) evaluating them based on scoring and adherence prediction capabilities. Results: Excluding the biased samples at the first time step, the predictive adherence accuracy of the SLAC models is from 60\% to 72\%, and for LSTM models, it is 66\% to 84\%, varying according to the time steps. The range of Root Mean Square Error (RMSE) for SLAC models is between 0.93 and 2.22, while for LSTM models it is between 1.09 and 1.77. Notably, these RMSEs are significantly lower than the random prediction error of 4.55. Conclusion: We creatively apply sequential models in the long-term management of SCIT with promising accuracy in the prediction of SCIT nonadherence in AR patients. While LSTM outperforms SLAC in adherence prediction, SLAC excels in score prediction for patients undergoing SCIT for AR. The state-action-based SLAC adds flexibility, presenting a novel and effective approach for managing long-term AIT., Comment: Frontiers in Pharmacology, research topic: Methods and Metrics to Measure Medication Adherence

Recently, reputation-based indirect reciprocity has been widely applied to the study on fairness behavior. Previous works mainly investigate indirect reciprocity by considering compulsory participation. While in reality, individuals may choose voluntary participation according to the opponent's reputation. It is still unclear how such reputation-based voluntary participation influences the evolution of fairness. To address this question, we introduce indirect reciprocity with voluntary participation into the dictator game (DG). We respectively consider good dictators or recipients can voluntarily participate in games when the opponents are assessed as bad. We theoretically calculate the fairness level under all social norms of third-order information. Our findings reveal that several social norms induce the high fairness level in both scenarios. However, more social norms lead to a high fairness level for voluntary participation of recipients, compared with the one of good dictators. The results also hold when the probability of voluntary participation is not low. Our results demonstrate that recipients' voluntary participation is more effective in promoting the emergence of fairness behavior.

Recent approaches such as ControlNet offer users fine-grained spatial control over text-to-image (T2I) diffusion models. However, auxiliary modules have to be trained for each type of spatial condition, model architecture, and checkpoint, putting them at odds with the diverse intents and preferences a human designer would like to convey to the AI models during the content creation process. In this work, we present FreeControl, a training-free approach for controllable T2I generation that supports multiple conditions, architectures, and checkpoints simultaneously. FreeControl designs structure guidance to facilitate the structure alignment with a guidance image, and appearance guidance to enable the appearance sharing between images generated using the same seed. Extensive qualitative and quantitative experiments demonstrate the superior performance of FreeControl across a variety of pre-trained T2I models. In particular, FreeControl facilitates convenient training-free control over many different architectures and checkpoints, allows the challenging input conditions on which most of the existing training-free methods fail, and achieves competitive synthesis quality with training-based approaches., Comment: Project Page: https://genforce.github.io/freecontrol/

The small-scale linear information in galaxy samples typically lost during non-linear growth can be restored to a certain level by the density field reconstruction, which has been demonstrated for improving the precision of the baryon acoustic oscillations (BAO) measurements. As proposed in the literature, a joint analysis of the power spectrum before and after the reconstruction enables an efficient extraction of information carried by high-order statistics. However, the statistics of the post-reconstruction density field are difficult to model. In this work, we circumvent this issue by developing an accurate emulator for the pre-reconstructed, post-reconstructed, and cross power spectra ($P_{\rm pre}$, $P_{\rm post}$, $P_{\rm cross}$) up to $k=0.5~h~{\rm Mpc^{-1}}$ based on the \textsc{Dark Quest} N-body simulations. The accuracy of the emulator is at percent level, namely, the error of the emulated monopole and quadrupole of the power spectra is less than $1\%$ and $10\%$ of the ground truth, respectively. A fit to an example power spectra using the emulator shows that the constraints on cosmological parameters get largely improved using $P_{\rm pre}$+$P_{\rm post}$+$P_{\rm cross}$ with $k_{\rm max}=0.25~h~{\rm Mpc^{-1}}$, compared to that derived from $P_{\rm pre}$ alone, namely, the constraints on ($\Omega_m$, $H_0$, $\sigma_8$) are tightened by $\sim41 \%-55\%$, and the uncertainties of the derived BAO and RSD parameters ($\alpha_{\perp}$, $\alpha_{||}$, $f\sigma_8$) shrink by $\sim 28\%-54\%$, respectively. This highlights the complementarity among $P_{\rm pre}$, $P_{\rm post}$ and $P_{\rm cross}$, which demonstrates the efficiency and practicability of a joint $P_{\rm pre}$, $P_{\rm post}$ and $P_{\rm cross}$ analysis for cosmological implications., Comment: 19 pages, 11 figures, 2 tables; accepted for publication in ApJ

The mass ($M_{\rm TOV}$) and radius ($R_{\rm TOV}$) of the maximum-mass nonrotating neutron star (NS) play a crucial role in constraining the elusive equation of state of cold dense matter and in predicting the fate of remnants from binary neutron star (BNS) mergers. In this study, we introduce a novel method to deduce these parameters by examining the mergers of second-generation (2G) black holes (BHs) with NSs. These 2G BHs are assumed to originate from supramassive neutron stars (SMNSs) formed in BNS mergers. Since the properties of the remnant BHs arising from the collapse of SMNSs follow a universal relation governed by $M_{\rm TOV}$ and $R_{\rm TOV}$, we anticipate that by analyzing a series ($\sim 100$ detections) of mass and spin measurements of the 2G BHs using the third-generation ground-based gravitational-wave detectors, $M_{\rm TOV}$ and $R_{\rm TOV}$ can be determined with a precision of $\sim 0.01M_\odot$ and $\sim 0.6$ km, respectively., Comment: 11 pages, 5 figures, ApJ published

Upcoming imaging surveys will use weak gravitational lensing to study the large-scale structure of the Universe, demanding sub-percent accuracy for precise cosmic shear measurements. We present a new differentiable implementation of our perturbation-based shear estimator (FPFS), using JAX, which is publicly available as part of a new suite of analytic shear algorithms called AnaCal. This code can analytically calibrate the shear response of any nonlinear observable constructed with the FPFS shapelets and detection modes utilizing auto-differentiation (AD), generalizing the formalism to include a family of shear estimators with corrections for detection and selection biases. Using the AD capability of JAX, it calculates the full Hessian matrix of the non-linear observables, which improves the previously presented second-order noise bias correction in the shear estimation. As an illustration of the power of the new AnaCal framework, we optimize the effective galaxy number density in the space of the generalized shear estimators using an LSST-like galaxy image simulation for the ten-year LSST. For the generic shear estimator, the magnitude of the multiplicative bias $|m|$ is below $3\times 10^{-3}$ (99.7% confidence interval), and the effective galaxy number density is improved by 5%. We also discuss some planned future additions to the AnaCal software suite to extend its applicability beyond the FPFS measurements., Comment: 9 pages, 7 figures, Accepted for publication in MNRAS

Vision Transformers achieve impressive accuracy across a range of visual recognition tasks. Unfortunately, their accuracy frequently comes with high computational costs. This is a particular issue in video recognition, where models are often applied repeatedly across frames or temporal chunks. In this work, we exploit temporal redundancy between subsequent inputs to reduce the cost of Transformers for video processing. We describe a method for identifying and re-processing only those tokens that have changed significantly over time. Our proposed family of models, Eventful Transformers, can be converted from existing Transformers (often without any re-training) and give adaptive control over the compute cost at runtime. We evaluate our method on large-scale datasets for video object detection (ImageNet VID) and action recognition (EPIC-Kitchens 100). Our approach leads to significant computational savings (on the order of 2-4x) with only minor reductions in accuracy., Comment: ICCV 2023

Deep neural networks have been widely used in various downstream tasks, especially those safety-critical scenario such as autonomous driving, but deep networks are often threatened by adversarial samples. Such adversarial attacks can be invisible to human eyes, but can lead to DNN misclassification, and often exhibits transferability between deep learning and machine learning models and real-world achievability. Adversarial attacks can be divided into white-box attacks, for which the attacker knows the parameters and gradient of the model, and black-box attacks, for the latter, the attacker can only obtain the input and output of the model. In terms of the attacker's purpose, it can be divided into targeted attacks and non-targeted attacks, which means that the attacker wants the model to misclassify the original sample into the specified class, which is more practical, while the non-targeted attack just needs to make the model misclassify the sample. The black box setting is a scenario we will encounter in practice.

Given a closed, oriented Lagrangian submanifold $L$ in a Liouville domain $\overline{M}$, one can define a Maurer-Cartan element with respect to a certain $L_\infty$-structure on the string homology $\widehat{H}_\ast^{S^1}(\mathcal{L}L;\mathbb{R})$, completed with respect to the action filtration. When the first Gutt-Hutchings capacity of $\overline{M}$ is finite, and $L$ is a $K(\pi,1)$ space, it leads to interesting geometric implications. In particular, we show that $L$ bounds a non-constant pseudoholomorphic disc of Maslov index 2. This confirms a general form of Audin's conjecture and generalizes the works of Fukaya and Irie in the case of $\mathbb{C}^n$ to a wide class of Liouville manifolds. In particular, when $\dim_\mathbb{R}(\overline{M})=6$, every closed, orientable, prime Lagrangian 3-manifold $L\subset\overline{M}$ is diffeomorphic to a spherical space form, or $S^1\times\Sigma_g$, where $\Sigma_g$ is a closed oriented surface., Comment: 76 pages, 6 figures. v3: minor revision, corrected a few typos and misattributions

Recently, hybrid bias expansions have emerged as a powerful approach to modelling the way in which galaxies are distributed in the Universe. Similarly, field-level emulators have recently become possible thanks to advances in machine learning and $N$-body simulations. In this paper we explore whether both techniques can be combined to provide a field-level model for the clustering of galaxies in real and redshift space. Specifically, here we will demonstrate that field-level emulators are able to accurately predict all the operators of a $2^{\rm nd}$-order hybrid bias expansion. The precision achieved in real and redshift space is similar to that obtained for the nonlinear matter power spectrum. This translates to roughly 1-2\% precision for the power spectrum of a BOSS and a Euclid-like galaxy sample up to $k\sim 0.6 h^{-1}$Mpc. Remarkably, this combined approach also delivers precise predictions for field-level galaxy statistics. Despite all these promising results, we detect several areas where further improvements are required. Therefore, this work serves as a road-map for the developments required for a more complete exploitation of upcoming large-scale structure surveys., Comment: 15 pages, 9 figures

This report describes our submission to the Ego4D Moment Queries Challenge 2023. Our submission extends ActionFormer, a latest method for temporal action localization. Our extension combines an improved ground-truth assignment strategy during training and a refined version of SoftNMS at inference time. Our solution is ranked 2nd on the public leaderboard with 26.62% average mAP and 45.69% Recall@1x at tIoU=0.5 on the test set, significantly outperforming the strong baseline from 2023 challenge. Our code is available at https://github.com/happyharrycn/actionformer_release.

The surging popularity of home assistants and their voice user interface (VUI) have made them an ideal central control hub for smart home devices. However, current form factors heavily rely on VUI, which poses accessibility and usability issues; some latest ones are equipped with additional cameras and displays, which are costly and raise privacy concerns. These concerns jointly motivate Beyond-Voice, a novel high-fidelity acoustic sensing system that allows commodity home assistant devices to track and reconstruct hand poses continuously. It transforms the home assistant into an active sonar system using its existing onboard microphones and speakers. We feed a high-resolution range profile to the deep learning model that can analyze the motions of multiple body parts and predict the 3D positions of 21 finger joints, bringing the granularity for acoustic hand tracking to the next level. It operates across different environments and users without the need for personalized training data. A user study with 11 participants in 3 different environments shows that Beyond-Voice can track joints with an average mean absolute error of 16.47mm without any training data provided by the testing subject., Comment: Accepted by IPSN 2024

Understanding the nonlinear relation between the shapes of halos or galaxies and the surrounding matter distribution is essential in accurate modeling of their intrinsic alignments. In the perturbative treatment, such nonlinear relation of the intrinsic alignments appears as higher-order shape bias parameters. In this paper, we present accurate measurements of the quadratic shape bias parameters by combining the \emph{full three-dimensional} power spectrum of the intrinsic alignments (i.e., without any projection) with the quadratic field method. In order to benefit from the full three-dimensional power spectrum we employ the spherical tensor decomposition of the three-dimensional shape field and measure their power spectra for the first time. In particular, we detect the vector and tensor power spectra in this basis, which cannot be explained by the widely-used nonlinear alignment model. Further, by cross-correlating the three-dimensional halo shape field with the quadratic shape bias operators from the initial condition of the same simulation to cancel cosmic variance, we effectively extract bispectrum information and detect quadratic shape bias parameters in the intrinsic alignments with high significance for the first time. We also compare these measurements with the prediction where quadratic shape biases are dynamically generated from the linear Lagrangian shape bias through the large-scale bulk flow. We find general agreement for all three biases with small deviations, which in practice could be negligible for the current photometric surveys. This implies that the advection prediction for the higher-order shape biases can be used as a prior in the cosmological analyses of intrinsic alignments., Comment: 19 pages, 5 figures; Minor changes to match version accepted for publication in JCAP