arXiv subjects

Yun Zhao

Publications and source records attributed to Yun Zhao.

At least 55 records · Page 3Linked to original sources

arXiv2023

Gemini in Reasoning: Unveiling Commonsense in Multimodal Large Language Models

The burgeoning interest in Multimodal Large Language Models (MLLMs), such as OpenAI's GPT-4V(ision), has significantly impacted both academic and industrial realms. These models enhance Large Language Models (LLMs) with advanced visual understanding capabilities, facilitating their application in a variety of multimodal tasks. Recently, Google introduced Gemini, a cutting-edge MLLM designed specifically for multimodal integration. Despite its advancements, preliminary benchmarks indicate that Gemini lags behind GPT models in commonsense reasoning tasks. However, this assessment, based on a limited dataset (i.e., HellaSWAG), does not fully capture Gemini's authentic commonsense reasoning potential. To address this gap, our study undertakes a thorough evaluation of Gemini's performance in complex reasoning tasks that necessitate the integration of commonsense knowledge across modalities. We carry out a comprehensive analysis of 12 commonsense reasoning datasets, ranging from general to domain-specific tasks. This includes 11 datasets focused solely on language, as well as one that incorporates multimodal elements. Our experiments across four LLMs and two MLLMs demonstrate Gemini's competitive commonsense reasoning capabilities. Additionally, we identify common challenges faced by current LLMs and MLLMs in addressing commonsense problems, underscoring the need for further advancements in enhancing the commonsense reasoning abilities of these models.

cs.CL↗

arXiv2023

Advanced Volleyball Stats for All Levels: Automatic Setting Tactic Detection and Classification with a Single Camera

This paper presents PathFinder and PathFinderPlus, two novel end-to-end computer vision frameworks designed specifically for advanced setting strategy classification in volleyball matches from a single camera view. Our frameworks combine setting ball trajectory recognition with a novel set trajectory classifier to generate comprehensive and advanced statistical data. This approach offers a fresh perspective for in-game analysis and surpasses the current level of granularity in volleyball statistics. In comparison to existing methods used in our baseline PathFinder framework, our proposed ball trajectory detection methodology in PathFinderPlus exhibits superior performance for classifying setting tactics under various game conditions. This robustness is particularly advantageous in handling complex game situations and accommodating different camera angles. Additionally, our study introduces an innovative algorithm for automatic identification of the opposing team's right-side (opposite) hitter's current row (front or back) during gameplay, providing critical insights for tactical analysis. The successful demonstration of our single-camera system's feasibility and benefits makes high-level technical analysis accessible to volleyball enthusiasts of all skill levels and resource availability. Furthermore, the computational efficiency of our system allows for real-time deployment, enabling in-game strategy analysis and on-the-spot gameplan adjustments.

cs.CV↗

arXiv2023

Are Large Language Models Ready for Healthcare? A Comparative Study on Clinical Language Understanding

Large language models (LLMs) have made significant progress in various domains, including healthcare. However, the specialized nature of clinical language understanding tasks presents unique challenges and limitations that warrant further investigation. In this study, we conduct a comprehensive evaluation of state-of-the-art LLMs, namely GPT-3.5, GPT-4, and Bard, within the realm of clinical language understanding tasks. These tasks span a diverse range, including named entity recognition, relation extraction, natural language inference, semantic textual similarity, document classification, and question-answering. We also introduce a novel prompting strategy, self-questioning prompting (SQP), tailored to enhance LLMs' performance by eliciting informative questions and answers pertinent to the clinical scenarios at hand. Our evaluation underscores the significance of task-specific learning strategies and prompting techniques for improving LLMs' effectiveness in healthcare-related tasks. Additionally, our in-depth error analysis on the challenging relation extraction task offers valuable insights into error distribution and potential avenues for improvement using SQP. Our study sheds light on the practical implications of employing LLMs in the specialized domain of healthcare, serving as a foundation for future research and the development of potential applications in healthcare settings.

cs.CL↗

arXiv2023

Characterization of SRB Measures for Random Dynamical Systems in a Banach space

This paper considers $C^2$ random dynamical systems in a Banach space, and proves that under some mild conditions, SRB measures are characterized by invariant measures satisfying Pesin's entropy formula, in which entropy is equal to the sum of positive Lyapunov exponents of the system. This can be regarded as a random version of the main result in Blumenthal and Young's paper \cite{Young17}.

math.DS↗

arXiv2023

An Empirical Study on the Robustness of the Segment Anything Model (SAM)

The Segment Anything Model (SAM) is a foundation model for general image segmentation. Although it exhibits impressive performance predominantly on natural images, understanding its robustness against various image perturbations and domains is critical for real-world applications where such challenges frequently arise. In this study we conduct a comprehensive robustness investigation of SAM under diverse real-world conditions. Our experiments encompass a wide range of image perturbations. Our experimental results demonstrate that SAM's performance generally declines under perturbed images, with varying degrees of vulnerability across different perturbations. By customizing prompting techniques and leveraging domain knowledge based on the unique characteristics of each dataset, the model's resilience to these perturbations can be enhanced, addressing dataset-specific challenges. This work sheds light on the limitations and strengths of SAM in real-world applications, promoting the development of more robust and versatile image segmentation solutions.

cs.CV↗

arXiv2023

All-optical frequency division on-chip using a single laser

The generation of spectrally pure high-frequency microwave signals is a critical functionality in fundamental and applied sciences, including metrology and communications. The development of optical frequency combs has enabled the powerful technique of optical frequency division (OFD) to produce microwave oscillations of the highest quality. The approaches for OFD demonstrated to date demand multiple lasers with space- and energy-consuming optical stabilization and electronic feedback components, resulting in device footprints incompatible with integration into a compact and robust photonic platform. Here, we demonstrate all-optical OFD on a single photonic chip driven with a single continuous-wave laser. We generate a dual-point frequency reference using the beat frequency of the signal and idler fields from a microresonator-based optical parametric oscillator (OPO), which achieves high phase stability due to the inherently strong signal-idler frequency correlations. We implement OFD by optically injecting the signal and idler fields from the OPO to a Kerr-comb microresonator on the same chip. We show that the two distinct dynamical states of Kerr cavities can be passively synchronized, allowing broadband frequency locking of the comb state, which transfers the stability of the OPO frequencies to the repetition rate of the Kerr comb. A 630-fold phase-noise reduction is observed when the Kerr comb is synchronized to the OPO, which represents the lowest noise generated on the silicon-nitride platform. Our work demonstrates a simple, effective approach for performing OFD and provides a pathway toward chip-scale devices that can generate microwave frequencies comparable to the purest tones produced in metrological laboratories. This technology can significantly boost the further development of data communications and microwave sensing.

physics.optics↗

arXiv2023

Large regenerative parametric amplification on chip at ultra-low pump powers

Chip-based optical amplifiers can significantly expand the functionalities of photonic devices. In particular, optical-parametric amplifiers (OPAs), with engineerable gain-spectra, are well-suited for nonlinear-photonic applications. Chip-based OPAs typically require long waveguides that occupy a large footprint, and high pump powers that cannot be easily produced with chip-scale lasers. We theoretically and experimentally demonstrate a microresonator-assisted regenerative OPA that benefits from the large nonlinearity enhancement of microresonators and yields a high gain in a small footprint. We achieve 30-dB parametric gain with only 9 mW of cw-pump power and show that the gain spectrum can be engineered to cover telecom channels inaccessible with Er-based amplifiers. We further demonstrate the amplification of Kerr-soliton comb lines and the preservation of their phase properties. Additionally, we demonstrate amplification by injection locking of optical-parametric oscillators, which corresponds to a regenerative amplifier pumped above the oscillation threshold. Novel dispersion engineering techniques such as coupled cavities and higher-order-dispersion phase matching can further extend the tunability and spectral coverage of our amplification schemes. The combination of high gain, small footprint, low pump power, and flexible gain-spectra engineering of our regenerative OPA is ideal for amplifying signals from the nanowatt to microwatt regimes for portable or space-based devices where ultralow electrical power levels are required and can lead to important applications in on-chip optical- and microwave-frequency synthesis and precise timekeeping.

physics.optics↗

arXiv2023

Brain Model State Space Reconstruction Using an LSTM Neural Network

Objective Kalman filtering has previously been applied to track neural model states and parameters, particularly at the scale relevant to EEG. However, this approach lacks a reliable method to determine the initial filter conditions and assumes that the distribution of states remains Gaussian. This study presents an alternative, data-driven method to track the states and parameters of neural mass models (NMMs) from EEG recordings using deep learning techniques, specifically an LSTM neural network. Approach An LSTM filter was trained on simulated EEG data generated by a neural mass model using a wide range of parameters. With an appropriately customised loss function, the LSTM filter can learn the behaviour of NMMs. As a result, it can output the state vector and parameters of NMMs given observation data as the input. Main Results Test results using simulated data yielded correlations with R squared of around 0.99 and verified that the method is robust to noise and can be more accurate than a nonlinear Kalman filter when the initial conditions of the Kalman filter are not accurate. As an example of real-world application, the LSTM filter was also applied to real EEG data that included epileptic seizures, and revealed changes in connectivity strength parameters at the beginnings of seizures. Significance Tracking the state vector and parameters of mathematical brain models is of great importance in the area of brain modelling, monitoring, imaging and control. This approach has no need to specify the initial state vector and parameters, which is very difficult to do in practice because many of the variables being estimated cannot be measured directly in physiological experiments. This method may be applied using any neural mass model and, therefore, provides a general, novel, efficient approach to estimate brain model variables that are often difficult to measure.

cs.LG↗

arXiv2023

Dimension approximation in smooth dynamical systems

For a non-conformal repeller $Λ$ of a $C^{1+α}$ map $f$ preserving an ergodic measure $μ$ of positive entropy, this paper shows that the Lyapunov dimension of $μ$ can be approximated gradually by the Carathéodory singular dimension of a sequence of horseshoes. For a $C^{1+α}$ diffeomorphism $f$ preserving a hyperbolic ergodic measure $μ$ of positive entropy, if $(f, μ)$ has only two Lyapunov exponents $λ_u(μ)>0>λ_s(μ)$, then the Hausdorff or lower box or upper box dimension of $μ$ can be approximated by the corresponding dimension of the horseshoes $\{Λ_n\}$. The same statement holds true if $f$ is a $C^1$ diffeomorphism with a dominated Oseledet's splitting with respect to $μ$.

math.DS↗

arXiv2022

Optical Neural Ordinary Differential Equations

Increasing the layer number of on-chip photonic neural networks (PNNs) is essential to improve its model performance. However, the successively cascading of network hidden layers results in larger integrated photonic chip areas. To address this issue, we propose the optical neural ordinary differential equations (ON-ODE) architecture that parameterizes the continuous dynamics of hidden layers with optical ODE solvers. The ON-ODE comprises the PNNs followed by the photonic integrator and optical feedback loop, which can be configured to represent residual neural networks (ResNet) and recurrent neural networks with effectively reduced chip area occupancy. For the interference-based optoelectronic nonlinear hidden layer, the numerical experiments demonstrate that the single hidden layer ON-ODE can achieve approximately the same accuracy as the two-layer optical ResNet in image classification tasks. Besides, the ONODE improves the model classification accuracy for the diffraction-based all-optical linear hidden layer. The time-dependent dynamics property of ON-ODE is further applied for trajectory prediction with high accuracy.

cs.LG↗

arXiv2022

Predicting the Need for Blood Transfusion in Intensive Care Units with Reinforcement Learning

As critically ill patients frequently develop anemia or coagulopathy, transfusion of blood products is a frequent intervention in the Intensive Care Units (ICU). However, inappropriate transfusion decisions made by physicians are often associated with increased risk of complications and higher hospital costs. In this work, we aim to develop a decision support tool that uses available patient information for transfusion decision-making on three common blood products (red blood cells, platelets, and fresh frozen plasma). To this end, we adopt an off-policy batch reinforcement learning (RL) algorithm, namely, discretized Batch Constrained Q-learning, to determine the best action (transfusion or not) given observed patient trajectories. Simultaneously, we consider different state representation approaches and reward design mechanisms to evaluate their impacts on policy learning. Experiments are conducted on two real-world critical care datasets: the MIMIC-III and the UCSF. Results demonstrate that policy recommendations on transfusion achieved comparable matching against true hospital policies via accuracy and weighted importance sampling evaluations on the MIMIC-III dataset. Furthermore, a combination of transfer learning (TL) and RL on the data-scarce UCSF dataset can provide up to $17.02% improvement in terms of accuracy, and up to 18.94% and 21.63% improvement in jump-start and asymptotic performance in terms of weighted importance sampling averaged over three transfusion tasks. Finally, simulations on transfusion decisions suggest that the transferred RL policy could reduce patients' estimated 28-day mortality rate by 2.74% and decreased acuity rate by 1.18% on the UCSF dataset.

cs.LG↗

arXiv2022

Morse-STF: Improved Protocols for Privacy-Preserving Machine Learning

Secure multi-party computation enables multiple mutually distrusting parties to perform computations on data without revealing the data itself, and has become one of the core technologies behind privacy-preserving machine learning. In this work, we present several improved privacy-preserving protocols for both linear and non-linear layers in machine learning. For linear layers, we present an extended beaver triple protocol for bilinear maps that significantly reduces communication of convolution layer. For non-linear layers, we introduce novel protocols for computing the sigmoid and softmax function. Both functions are essential building blocks for machine learning training of classification tasks. Our protocols are both more scalable and robust than prior constructions, and improves runtime performance by 3-17x. Finally, we introduce Morse-STF, an end-to-end privacy-preserving system for machine learning training that leverages all these improved protocols. Our system achieves a 1.8x speedup on logistic regression and 3.9-4.9x speedup on convolutional neural networks compared to prior state-of-the-art systems.

cs.CR↗

arXiv2022

Hölder continuity of Oseledets subspaces for linear cocycles on Banach spaces

Let $f:X\to X$ be an invertible Lipschitz transformation on a compact metric space $X$. Given a Hölder continuous invertible operator cocycles on a Banach space and an $f$-invariant ergodic measure, this paper establishes the Hölder continuity of Oseledets subspaces over a compact set of arbitrarily large measure. This extends a result in \cite{Simion16} for invertible operator cocycles on a Banach space. Finally, this paper proves the Hölder continuity in the non-invertible case.

math.DS↗

arXiv2022

Integrating Physiological Time Series and Clinical Notes with Transformer for Early Prediction of Sepsis

Sepsis is a leading cause of death in the Intensive Care Units (ICU). Early detection of sepsis is critical for patient survival. In this paper, we propose a multimodal Transformer model for early sepsis prediction, using the physiological time series data and clinical notes for each patient within $36$ hours of ICU admission. Specifically, we aim to predict sepsis using only the first 12, 18, 24, 30 and 36 hours of laboratory measurements, vital signs, patient demographics, and clinical notes. We evaluate our model on two large critical care datasets: MIMIC-III and eICU-CRD. The proposed method is compared with six baselines. In addition, ablation analysis and case studies are conducted to study the influence of each individual component of the model and the contribution of each data modality for early sepsis prediction. Experimental results demonstrate the effectiveness of our method, which outperforms competitive baselines on all metrics.

cs.LG↗

arXiv2022

Enhancing Transformer Efficiency for Multivariate Time Series Classification

Most current multivariate time series (MTS) classification algorithms focus on improving the predictive accuracy. However, for large-scale (either high-dimensional or long-sequential) time series (TS) datasets, there is an additional consideration: to design an efficient network architecture to reduce computational costs such as training time and memory footprint. In this work we propose a methodology based on module-wise pruning and Pareto analysis to investigate the relationship between model efficiency and accuracy, as well as its complexity. Comprehensive experiments on benchmark MTS datasets illustrate the effectiveness of our method.

cs.LG↗

arXiv2021

Empirical Quantitative Analysis of COVID-19 Forecasting Models

COVID-19 has been a public health emergency of international concern since early 2020. Reliable forecasting is critical to diminish the impact of this disease. To date, a large number of different forecasting models have been proposed, mainly including statistical models, compartmental models, and deep learning models. However, due to various uncertain factors across different regions such as economics and government policy, no forecasting model appears to be the best for all scenarios. In this paper, we perform quantitative analysis of COVID-19 forecasting of confirmed cases and deaths across different regions in the United States with different forecasting horizons, and evaluate the relative impacts of the following three dimensions on the predictive performance (improvement and variation) through different evaluation metrics: model selection, hyperparameter tuning, and the length of time series required for training. We find that if a dimension brings about higher performance gains, if not well-tuned, it may also lead to harsher performance penalties. Furthermore, model selection is the dominant factor in determining the predictive performance. It is responsible for both the largest improvement and the largest variation in performance in all prediction tasks across different regions. While practitioners may perform more complicated time series analysis in practice, they should be able to achieve reasonable results if they have adequate insight into key decisions like model selection.

cs.LG↗

arXiv2021

Theory of $χ^{(2)}$-microresonator-based frequency conversion

Microresonator-based platforms with $χ^{(2)}$ nonlinearities have the potential to perform frequency conversion at high efficiencies and ultralow powers with small footprints. The standard doctrine for achieving high conversion efficiency in cavity-based devices requires "perfect matching", that is, zero phase mismatch while all relevant frequencies are precisely at a cavity resonance, which is difficult to achieve in integrated platforms due to fabrication errors and limited tunabilities. In this Letter, we show that the violation of perfect matching does not necessitate a reduction in conversion efficiency. On the contrary, in many cases, mismatches should be intentionally introduced to improve the efficiency or tunability of conversion. We identify the universal conditions for maximizing the efficiency of cavity-based frequency conversion and show a straightforward approach to fully compensate for parasitic processes such as thermorefractive and photorefractive effects that, typically, can limit the conversion efficiency. We also show rigorously that these high-efficiency states are stable.

physics.optics↗

arXiv2021

Multiple Organ Failure Prediction with Classifier-Guided Generative Adversarial Imputation Networks

Multiple organ failure (MOF) is a severe syndrome with a high mortality rate among Intensive Care Unit (ICU) patients. Early and precise detection is critical for clinicians to make timely decisions. An essential challenge in applying machine learning models to electronic health records (EHRs) is the pervasiveness of missing values. Most existing imputation methods are involved in the data preprocessing phase, failing to capture the relationship between data and outcome for downstream predictions. In this paper, we propose classifier-guided generative adversarial imputation networks Classifier-GAIN) for MOF prediction to bridge this gap, by incorporating both observed data and label information. Specifically, the classifier takes imputed values from the generator(imputer) to predict task outcomes and provides additional supervision signals to the generator by joint training. The classifier-guide generator imputes missing values with label-awareness during training, improving the classifier's performance during inference. We conduct extensive experiments showing that our approach consistently outperforms classical and state-of-art neural baselines across a range of missing data scenarios and evaluation metrics.

cs.LG↗