Searcharxiv⌕ Search

arXiv subjects

Qi Yu

Publications and source records attributed to Qi Yu.

At least 73 records · Page 4Linked to original sources

Deep Reinforced Attention Regression for Partial Sketch Based Image Retrieval

Fine-Grained Sketch-Based Image Retrieval (FG-SBIR) aims at finding a specific image from a large gallery given a query sketch. Despite the widespread applicability of FG-SBIR in many critical domains (e.g., crime activity tracking), existing approaches still suffer from a low accuracy while being sensitive to external noises such as unnecessary strokes in the sketch. The retrieval performance will further deteriorate under a more practical on-the-fly setting, where only a partially complete sketch with only a few (noisy) strokes are available to retrieve corresponding images. We propose a novel framework that leverages a uniquely designed deep reinforcement learning model that performs a dual-level exploration to deal with partial sketch training and attention region selection. By enforcing the model's attention on the important regions of the original sketches, it remains robust to unnecessary stroke noises and improve the retrieval accuracy by a large margin. To sufficiently explore partial sketches and locate the important regions to attend, the model performs bootstrapped policy gradient for global exploration while adjusting a standard deviation term that governs a locator network for local exploration. The training process is guided by a hybrid loss that integrates a reinforcement loss and a supervised loss. A dynamic ranking reward is developed to fit the on-the-fly image retrieval process using partial sketches. The extensive experimentation performed on three public datasets shows that our proposed approach achieves the state-of-the-art performance on partial sketch based image retrieval.

cs.CV↗

Simultaneous estimation of parameters and the state of an optical parametric oscillator system

In this paper, we consider the filtering problem of an optical parametric oscillator (OPO). The OPO pump power may fluctuate due to environmental disturbances, resulting in uncertainty in the system modeling. Thus, both the state and the unknown parameter may need to be estimated simultaneously. We formulate this problem using a state-space representation of the OPO dynamics. Under the assumption of Gaussianity and proper constraints, the dual Kalman filter method and the joint extended Kalman filter method are employed to simultaneously estimate the system state and the pump power. Numerical examples demonstrate the effectiveness of the proposed algorithms.

eess.SY↗

DRIVE: Deep Reinforced Accident Anticipation with Visual Explanation

Traffic accident anticipation aims to accurately and promptly predict the occurrence of a future accident from dashcam videos, which is vital for a safety-guaranteed self-driving system. To encourage an early and accurate decision, existing approaches typically focus on capturing the cues of spatial and temporal context before a future accident occurs. However, their decision-making lacks visual explanation and ignores the dynamic interaction with the environment. In this paper, we propose Deep ReInforced accident anticipation with Visual Explanation, named DRIVE. The method simulates both the bottom-up and top-down visual attention mechanism in a dashcam observation environment so that the decision from the proposed stochastic multi-task agent can be visually explained by attentive regions. Moreover, the proposed dense anticipation reward and sparse fixation reward are effective in training the DRIVE model with our improved reinforcement learning algorithm. Experimental results show that the DRIVE model achieves state-of-the-art performance on multiple real-world traffic accident datasets. Code and pre-trained model are available at \url{https://www.rit.edu/actionlab/drive}.

cs.CV↗

Evidential Deep Learning for Open Set Action Recognition

In a real-world scenario, human actions are typically out of the distribution from training data, which requires a model to both recognize the known actions and reject the unknown. Different from image data, video actions are more challenging to be recognized in an open-set setting due to the uncertain temporal dynamics and static bias of human actions. In this paper, we propose a Deep Evidential Action Recognition (DEAR) method to recognize actions in an open testing set. Specifically, we formulate the action recognition problem from the evidential deep learning (EDL) perspective and propose a novel model calibration method to regularize the EDL training. Besides, to mitigate the static bias of video representation, we propose a plug-and-play module to debias the learned representation through contrastive learning. Experimental results show that our DEAR method achieves consistent performance gain on multiple mainstream action recognition models and benchmarks. Code and pre-trained models are available at {\small{\url{https://www.rit.edu/actionlab/dear}}}.

cs.CV↗

Comprehensive fabrication of SNAP microresonators by a femtosecond laser

Surface nanoscale axial photonics (SNAP) microresonators with nanoscale effective radius variation (ERV) along optical fiber axis can be fabricated by inscribing axially oriented lines inside the fiber with a femtosecond laser. The optimization of variable parameters in the femtosecond laser inscription technique is of great significance for flexible and ultra-high precision control of ERV, which is vital to the performance of SNAP devices. Here, we present the first systematical investigation on the relationships between the various controllable fabrication parameters and the introduced ERV of the SNAP microresonators. Specifically, both the qualitative and quantitative processing principles were revealed. As a proof-of-principle, by comprehensively optimizing the fabrication parameters, we realized a SNAP microresonator with the characteristics of both small axial size and maximal ERV, which is almost impossible to realize with other SNAP fabrication techniques. Our work promotes the fs laser inscription technology to be a flexible and versatile approach for fabricating the SNAP devices with ultra-high precision, ultra-low loss and high robustness.

physics.optics↗

Presenting and Evaluating the Impact of Experiential Learning in Computing Accessibility Education

Studies indicate that much of the software created today is not accessible to all users, indicating that developers don't see the need to devote sufficient resources to creating accessible software. Compounding this problem, there is a lack of robust, easily adoptable educational accessibility material available to instructors for inclusion in their curricula. To address these issues, we have created five Accessibility Learning Labs (ALL) using an experiential learning structure. The labs are designed to educate and create awareness of accessibility needs in computing. The labs enable easy classroom integration by providing instructors with complete educational materials including lecture slides, activities, and quizzes. The labs are hosted on our servers and require only a browser to be utilized. To demonstrate the benefit of our material and the potential benefits of our experiential lab format with empathy-creating material, we conducted a study involving 276 students in ten sections of an introductory computing course. Our findings include: (I) The demonstrated potential of the proposed experiential learning format and labs are effective in motivating and educating students about the importance of accessibility (II) The labs are effective in informing students about foundational accessibility topics (III) Empathy-creating material is demonstrated to be a beneficial component in computing accessibility education, supporting students in placing a higher value on the importance of creating accessible software. Created labs and project materials are publicly available on the project website: http://all.rit.edu

cs.SE↗

Uncertainty-based Traffic Accident Anticipation with Spatio-Temporal Relational Learning

Traffic accident anticipation aims to predict accidents from dashcam videos as early as possible, which is critical to safety-guaranteed self-driving systems. With cluttered traffic scenes and limited visual cues, it is of great challenge to predict how long there will be an accident from early observed frames. Most existing approaches are developed to learn features of accident-relevant agents for accident anticipation, while ignoring the features of their spatial and temporal relations. Besides, current deterministic deep neural networks could be overconfident in false predictions, leading to high risk of traffic accidents caused by self-driving systems. In this paper, we propose an uncertainty-based accident anticipation model with spatio-temporal relational learning. It sequentially predicts the probability of traffic accident occurrence with dashcam videos. Specifically, we propose to take advantage of graph convolution and recurrent networks for relational feature learning, and leverage Bayesian neural networks to address the intrinsic variability of latent relational representations. The derived uncertainty-based ranking loss is found to significantly boost model performance by improving the quality of relational features. In addition, we collect a new Car Crash Dataset (CCD) for traffic accident anticipation which contains environmental attributes and accident reasons annotations. Experimental results on both public and the newly-compiled datasets show state-of-the-art performance of our model. Our code and CCD dataset are available at https://github.com/Cogito2012/UString.

cs.CV↗

Object-Aware Centroid Voting for Monocular 3D Object Detection

Monocular 3D object detection aims to detect objects in a 3D physical world from a single camera. However, recent approaches either rely on expensive LiDAR devices, or resort to dense pixel-wise depth estimation that causes prohibitive computational cost. In this paper, we propose an end-to-end trainable monocular 3D object detector without learning the dense depth. Specifically, the grid coordinates of a 2D box are first projected back to 3D space with the pinhole model as 3D centroids proposals. Then, a novel object-aware voting approach is introduced, which considers both the region-wise appearance attention and the geometric projection distribution, to vote the 3D centroid proposals for 3D object localization. With the late fusion and the predicted 3D orientation and dimension, the 3D bounding boxes of objects can be detected from a single RGB image. The method is straightforward yet significantly superior to other monocular-based methods. Extensive experimental results on the challenging KITTI benchmark validate the effectiveness of the proposed method.

cs.CV↗

Generation of accessible sets in the dynamical modelling of quantum network systems

In this paper, we consider the dynamical modeling of a class of quantum network systems consisting of qubits. Qubit probes are employed to measure a set of selected nodes of the quantum network systems. For a variety of applications, a state space model is a useful way to model the system dynamics. To construct a state space model for a quantum network system, the major task is to find an accessible set containing all of the operators coupled to the measurement operators. This paper focuses on the generation of a proper accessible set for a given system and measurement scheme. We provide analytic results on simplifying the process of generating accessible sets for systems with a time-independent Hamiltonian. Since the order of elements in the accessible set determines the form of state space matrices, guidance is provided to effectively arrange the ordering of elements in the state vector. Defining a system state according to the accessible set, one can develop a state space model with a special pattern inherited from the system structure. As a demonstration, we specifically consider a typical 1D-chain system with several common measurements, and employ the proposed method to determine its accessible set.

eess.SY↗

Improving the Decision-Making Process of Self-Adaptive Systems by Accounting for Tactic Volatility

When self-adaptive systems encounter changes within their surrounding environments, they enact tactics to perform necessary adaptations. For example, a self-adaptive cloud-based system may have a tactic that initiates additional computing resources when response time thresholds are surpassed, or there may be a tactic to activate a specific security measure when an intrusion is detected. In real-world environments, these tactics frequently experience tactic volatility which is variable behavior during the execution of the tactic. Unfortunately, current self-adaptive approaches do not account for tactic volatility in their decision-making processes, and merely assume that tactics do not experience volatility. This limitation creates uncertainty in the decision-making process and may adversely impact the system's ability to effectively and efficiently adapt. Additionally, many processes do not properly account for volatility that may effect the system's Service Level Agreement (SLA). This can limit the system's ability to act proactively, especially when utilizing tactics that contain latency. To address the challenge of sufficiently accounting for tactic volatility, we propose a Tactic Volatility Aware (TVA) solution. Using Multiple Regression Analysis (MRA), TVA enables self-adaptive systems to accurately estimate the cost and time required to execute tactics. TVA also utilizes Autoregressive Integrated Moving Average (ARIMA) for time series forecasting, allowing the system to proactively maintain specifications.

cs.AI↗

Hybrid filtering for a class of nonlinear quantum systems subject to classical stochastic disturbances

A hybrid quantum-classical filtering problem, where a qubit system is disturbed by a classical stochastic process, is investigated. The strategy is to model the classical disturbance by using an optical cavity. Relations between classical disturbances and the cavity analog system are analyzed. The dynamics of the enlarged quantum network system, which includes a qubit system and a cavity system, are derived. A stochastic master equation for the qubit-cavity hybrid system is given, based on which estimates for the state of the cavity system and the classical signal are obtained. The quantum extended Kalman filter is employed to achieve efficient computation. Numerical results are presented to illustrate the effectiveness of our methods.

eess.SY↗

On the capability of a class of quantum sensors

Quantum sensors may provide extremely high sensitivity and precision to extract key information in a quantum or classical physical system. A fundamental question is whether a quantum sensor is capable of uniquely inferring unknown parameters in a system for a given structure of the quantum sensor and admissible measurement on the sensor. In this paper, we investigate the capability of a class of quantum sensors which consist of either a single qubit or two qubits. A quantum sensor is coupled to a spin chain system to extract information of unknown parameters in the system. With given initialisation and measurement schemes, we employ the similarity transformation approach and the Grobner basis method to prove that a single-qubit quantum sensor cannot effectively estimate the unknown parameters in the spin chain system while the two-qubit quantum sensor can. The work demonstrates that it is a feasible method to enhance the capability of quantum sensors by increasing the number of qubits in the quantum sensors for some practical applications.

quant-ph↗

Dictionary Learning with BLOTLESS Update

Algorithms for learning a dictionary to sparsely represent a given dataset typically alternate between sparse coding and dictionary update stages. Methods for dictionary update aim to minimise expansion error by updating dictionary vectors and expansion coefficients given patterns of non-zero coefficients obtained in the sparse coding stage. We propose a block total least squares (BLOTLESS) algorithm for dictionary update. BLOTLESS updates a block of dictionary elements and the corresponding sparse coefficients simultaneously. In the error free case, three necessary conditions for exact recovery are identified. Lower bounds on the number of training data are established so that the necessary conditions hold with high probability. Numerical simulations show that the bounds approximate well the number of training data needed for exact dictionary recovery. Numerical experiments further demonstrate several benefits of dictionary learning with BLOTLESS update compared with state-of-the-art algorithms especially when the amount of training data is small.

eess.SP↗

Rectangular SNAP microresonator fabricated with a femtosecond laser

SNAP microresonators, which are fabricated by nanoscale effective radius variation (ERV) of the optical fiber with sub-angstrom precision, can be potentially used as miniature classical and quantum signal processors, frequency comb generators, as well as ultraprecise microfluidic and environmental optical sensors. Many of these applications require the introduction of nanoscale ERV with a large contrast α which is defined as the maximum shift of the fiber cutoff wavelength introduced per unit length of the fiber axis. The previously developed fabrication methods of SNAP structures, which used focused CO2 and femtosecond laser beams, achieved α ~ 0.02 nm/um. Here we develop a new fabrication method of SNAP microresonators with a femtosecond laser which allows us to demonstrate a 50-fold improvement of previous results and achieve α ~ 1 nm/um. Furthermore, our fabrication method enables the introduction of ERV which is several times larger than the maximum ERV demonstrated previously. As an example, we fabricate a rectangular SNAP resonator and investigate its group delay characteristics. Our experimental results are in good agreement with theoretical simulations. Overall, the developed approach allows us to reduce the axial scale of SNAP structures by an order of magnitude.

physics.optics↗

edge2vec: Representation learning using edge semantics for biomedical knowledge discovery

Representation learning provides new and powerful graph analytical approaches and tools for the highly valued data science challenge of mining knowledge graphs. Since previous graph analytical methods have mostly focused on homogeneous graphs, an important current challenge is extending this methodology for richly heterogeneous graphs and knowledge domains. The biomedical sciences are such a domain, reflecting the complexity of biology, with entities such as genes, proteins, drugs, diseases, and phenotypes, and relationships such as gene co-expression, biochemical regulation, and biomolecular inhibition or activation. Therefore, the semantics of edges and nodes are critical for representation learning and knowledge discovery in real world biomedical problems. In this paper, we propose the edge2vec model, which represents graphs considering edge semantics. An edge-type transition matrix is trained by an Expectation-Maximization approach, and a stochastic gradient descent model is employed to learn node embedding on a heterogeneous graph via the trained transition matrix. edge2vec is validated on three biomedical domain tasks: biomedical entity classification, compound-gene bioactivity prediction, and biomedical information retrieval. Results show that by considering edge-types into node embedding learning in heterogeneous graphs, \textbf{edge2vec}\ significantly outperforms state-of-the-art models on all three tasks. We propose this method for its added value relative to existing graph analytical methodology, and in the real world context of biomedical knowledge discovery applicability.

cs.IR↗

Quantum speed-up in solving the maximal clique problem

The maximal clique problem, to find the maximally sized clique in a given graph, is classically an NP-complete computational problem, which has potential applications ranging from electrical engineering, computational chemistry, bioinformatics to social networks. Here we develop a quantum algorithm to solve the maximal clique problem for any graph $G$ with $n$ vertices with quadratic speed-up over its classical counterparts, where the time and spatial complexities are reduced to, respectively, $O(\sqrt{2^{n}})$ and $O(n^{2})$. With respect to oracle-related quantum algorithms for the NP-complete problems, we identify our algorithm to be optimal. To justify the feasibility of the proposed quantum algorithm, we have successfully solved an exemplified clique problem for a graph $G$ with two vertices and one edge by carrying out a nuclear magnetic resonance experiment involving four qubits.

quant-ph↗

Spectral analyses of trans- and cis-DOCO transients via comb spectroscopy

We use time-resolved direct frequency comb spectroscopy in the mid-infrared to obtain high-resolution rovibrational spectra of products produced from the OD+CO reaction. In this work, we present spectral analyses for isotopologues of the transient DOCO radicals from this reaction in the OD stretch region. The analyses were performed with aid of two different theoretical approaches based on both perturbation theory and variational calculations used for prediction of rovibrational spectra of polyatomic molecules. We discuss the advantages and challenges of our current approach for studying spectroscopy and dynamics of transient molecules.

physics.chem-ph↗

Hybrid Filtering for a Class of Quantum Systems with Classical Disturbances

A filtering problem for a class of quantum systems disturbed by a classical stochastic process is investigated in this paper. The classical disturbance process, which is assumed to be described by a linear stochastic differential equation, is modeled by a quantum cavity model. Then the hybrid quantum-classical system is described by a combined quantum system consisting of two quantum cavity subsystems. Quantum filtering theory and a quantum extended Kalman filter method are employed to estimate the states of the combined quantum system. An estimate of the classical stochastic process is derived from the estimate of the combined quantum system. The effectiveness and performance of the proposed methods are illustrated by numerical results.

quant-ph↗