Showing 1–46 of 46 results for author: To, T

Deep learning for automated detection of breast cancer in deep ultraviolet fluorescence images with diffusion probabilistic model

Authors: Sepehr Salem Ghahfarokhi, Tyrell To, Julie Jorns, Tina Yen, Bing Yu, Dong Hye Ye

Abstract: Data limitation is a significant challenge in applying deep learning to medical images. Recently, the diffusion probabilistic model (DPM) has shown the potential to generate high-quality images by converting Gaussian random noise into realistic images. In this paper, we apply the DPM to augment the deep ultraviolet fluorescence (DUV) image dataset with an aim to improve breast cancer classificatio… ▽ More Data limitation is a significant challenge in applying deep learning to medical images. Recently, the diffusion probabilistic model (DPM) has shown the potential to generate high-quality images by converting Gaussian random noise into realistic images. In this paper, we apply the DPM to augment the deep ultraviolet fluorescence (DUV) image dataset with an aim to improve breast cancer classification for intraoperative margin assessment. For classification, we divide the whole surface DUV image into small patches and extract convolutional features for each patch by utilizing the pre-trained ResNet. Then, we feed them into an XGBoost classifier for patch-level decisions and then fuse them with a regional importance map computed by Grad-CAM++ for whole surface-level prediction. Our experimental results show that augmenting the training dataset with the DPM significantly improves breast cancer detection performance in DUV images, increasing accuracy from 93% to 97%, compared to using Affine transformations and ProGAN. △ Less

Submitted 1 July, 2024; originally announced July 2024.

Comments: IEEE International Symposium on Biomedical Imaging 2024

Kähler families of Green's functions

Authors: Vincent Guedj, Tat Dat Tô

Abstract: In a remarkable series of works, Guo, Phong, Song, and Sturm have obtained key uniform estimates for the Green's functions associated with certain Kähler metrics. In this note, we broaden the scope of their techniques by removing one of their assumptions and allowing the complex structure to vary. We apply our results to various families of canonical Kähler metrics. In a remarkable series of works, Guo, Phong, Song, and Sturm have obtained key uniform estimates for the Green's functions associated with certain Kähler metrics. In this note, we broaden the scope of their techniques by removing one of their assumptions and allowing the complex structure to vary. We apply our results to various families of canonical Kähler metrics. △ Less

Submitted 2 June, 2024; v1 submitted 27 May, 2024; originally announced May 2024.

Comments: Minor changes, references added

ViWikiFC: Fact-Checking for Vietnamese Wikipedia-Based Textual Knowledge Source

Authors: Hung Tuan Le, Long Truong To, Manh Trong Nguyen, Kiet Van Nguyen

Abstract: Fact-checking is essential due to the explosion of misinformation in the media ecosystem. Although false information exists in every language and country, most research to solve the problem mainly concentrated on huge communities like English and Chinese. Low-resource languages like Vietnamese are necessary to explore corpora and models for fact verification. To bridge this gap, we construct ViWik… ▽ More Fact-checking is essential due to the explosion of misinformation in the media ecosystem. Although false information exists in every language and country, most research to solve the problem mainly concentrated on huge communities like English and Chinese. Low-resource languages like Vietnamese are necessary to explore corpora and models for fact verification. To bridge this gap, we construct ViWikiFC, the first manual annotated open-domain corpus for Vietnamese Wikipedia Fact Checking more than 20K claims generated by converting evidence sentences extracted from Wikipedia articles. We analyze our corpus through many linguistic aspects, from the new dependency rate, the new n-gram rate, and the new word rate. We conducted various experiments for Vietnamese fact-checking, including evidence retrieval and verdict prediction. BM25 and InfoXLM (Large) achieved the best results in two tasks, with BM25 achieving an accuracy of 88.30% for SUPPORTS, 86.93% for REFUTES, and only 56.67% for the NEI label in the evidence retrieval task, InfoXLM (Large) achieved an F1 score of 86.51%. Furthermore, we also conducted a pipeline approach, which only achieved a strict accuracy of 67.00% when using InfoXLM (Large) and BM25. These results demonstrate that our dataset is challenging for the Vietnamese language model in fact-checking tasks. △ Less

Submitted 13 May, 2024; originally announced May 2024.

Singular cscK metrics on smoothable varieties

Authors: Chung-Ming Pan, Tat Dat Tô, Antonio Trusiani

Abstract: We prove the lower semi-continuity of the coercivity threshold of Mabuchi functional along a degenerate family of normal compact Kähler varieties with klt singularities. Moreover, we establish the existence of singular cscK metrics on $\mathbb{Q}$-Gorenstein smoothable klt varieties when the Mabuchi functional is coercive, these arise as a limit of cscK metrics on close-by fibres. The proof relies… ▽ More We prove the lower semi-continuity of the coercivity threshold of Mabuchi functional along a degenerate family of normal compact Kähler varieties with klt singularities. Moreover, we establish the existence of singular cscK metrics on $\mathbb{Q}$-Gorenstein smoothable klt varieties when the Mabuchi functional is coercive, these arise as a limit of cscK metrics on close-by fibres. The proof relies on develo** a novel strong topology of pluripotential theory in families and establishing uniform estimates for cscK metrics. △ Less

Submitted 22 January, 2024; v1 submitted 21 December, 2023; originally announced December 2023.

Comments: 46 pages; v2: removed appendix, changed the proof of Theorem 5.4, corrected some typos

Raijū: Reinforcement Learning-Guided Post-Exploitation for Automating Security Assessment of Network Systems

Authors: Van-Hau Pham, Hien Do Hoang, Phan Thanh Trung, Van Dinh Quoc, Trong-Nghia To, Phan The Duy

Abstract: In order to assess the risks of a network system, it is important to investigate the behaviors of attackers after successful exploitation, which is called post-exploitation. Although there are various efficient tools supporting post-exploitation implementation, no application can automate this process. Most of the steps of this process are completed by experts who have profound knowledge of securi… ▽ More In order to assess the risks of a network system, it is important to investigate the behaviors of attackers after successful exploitation, which is called post-exploitation. Although there are various efficient tools supporting post-exploitation implementation, no application can automate this process. Most of the steps of this process are completed by experts who have profound knowledge of security, known as penetration testers or pen-testers. To this end, our study proposes the Raijū framework, a Reinforcement Learning (RL)-driven automation approach that assists pen-testers in quickly implementing the process of post-exploitation for security-level evaluation in network systems. We implement two RL algorithms, Advantage Actor-Critic (A2C) and Proximal Policy Optimization (PPO), to train specialized agents capable of making intelligent actions, which are Metasploit modules to automatically launch attacks of privileges escalation, gathering hashdump, and lateral movement. By leveraging RL, we aim to empower these agents with the ability to autonomously select and execute actions that can exploit vulnerabilities in target systems. This approach allows us to automate certain aspects of the penetration testing workflow, making it more efficient and responsive to emerging threats and vulnerabilities. The experiments are performed in four real environments with agents trained in thousands of episodes. The agents automatically select actions and launch attacks on the environments and achieve over 84\% of successful attacks with under 55 attack steps given. Moreover, the A2C algorithm has proved extremely effective in the selection of proper actions for automation of post-exploitation. △ Less

Submitted 27 September, 2023; originally announced September 2023.

On the Effectiveness of Adversarial Samples against Ensemble Learning-based Windows PE Malware Detectors

Authors: Trong-Nghia To, Danh Le Kim, Do Thi Thu Hien, Nghi Hoang Khoa, Hien Do Hoang, Phan The Duy, Van-Hau Pham

Abstract: Recently, there has been a growing focus and interest in applying machine learning (ML) to the field of cybersecurity, particularly in malware detection and prevention. Several research works on malware analysis have been proposed, offering promising results for both academic and practical applications. In these works, the use of Generative Adversarial Networks (GANs) or Reinforcement Learning (RL… ▽ More Recently, there has been a growing focus and interest in applying machine learning (ML) to the field of cybersecurity, particularly in malware detection and prevention. Several research works on malware analysis have been proposed, offering promising results for both academic and practical applications. In these works, the use of Generative Adversarial Networks (GANs) or Reinforcement Learning (RL) can aid malware creators in crafting metamorphic malware that evades antivirus software. In this study, we propose a mutation system to counteract ensemble learning-based detectors by combining GANs and an RL model, overcoming the limitations of the MalGAN model. Our proposed FeaGAN model is built based on MalGAN by incorporating an RL model called the Deep Q-network anti-malware Engines Attacking Framework (DQEAF). The RL model addresses three key challenges in performing adversarial attacks on Windows Portable Executable malware, including format preservation, executability preservation, and maliciousness preservation. In the FeaGAN model, ensemble learning is utilized to enhance the malware detector's evasion ability, with the generated adversarial patterns. The experimental results demonstrate that 100\% of the selected mutant samples preserve the format of executable files, while certain successes in both executability preservation and maliciousness preservation are achieved, reaching a stable success rate. △ Less

Submitted 24 September, 2023; originally announced September 2023.

Global harmonic analysis for $Φ^4_3$ on closed Riemannian manifolds

Authors: I. Bailleul, N. V. Dang, L. Ferdinand, T. D. Tô

Abstract: Following Parisi \& Wu's paradigm of stochastic quantization, we constructed in \cite{BDFT} a $Φ^4$ measure on an arbitrary closed, compact Riemannian manifold of dimension $3$ as an invariant measure of a singular stochastic partial differential equation. This solves a longstanding open problem in quantum fields on curved backgrounds. In the present work, we build all the harmonic and microlocal… ▽ More Following Parisi \& Wu's paradigm of stochastic quantization, we constructed in \cite{BDFT} a $Φ^4$ measure on an arbitrary closed, compact Riemannian manifold of dimension $3$ as an invariant measure of a singular stochastic partial differential equation. This solves a longstanding open problem in quantum fields on curved backgrounds. In the present work, we build all the harmonic and microlocal analysis tools that are needed in \cite{BDFT}. In particular, we extend the approach of Jagannath--Perkowski to the vectorial $Φ^4_3$ model by introducing a new Cole-Hopf transform involving random bundle maps. △ Less

Submitted 18 July, 2023; v1 submitted 13 June, 2023; originally announced June 2023.

Comments: references added, details added in Section 5.2, Section B added

MSC Class: 81T20; 35R60

$Φ^4_3$ measures on compact Riemannian $3$-manifolds

Authors: I. Bailleul, N. V. Dang, L. Ferdinand, T. D. Tô

Abstract: We construct the $Φ^4_3$ measure on an arbitrary 3-dimensional compact Riemannian manifold without boundary as an invariant probability measure of a singular stochastic partial differential equation. Proving the nontriviality and the covariance under Riemannian isometries of that measure gives for the first time a non-perturbative, non-topological interacting Euclidean quantum field theory on curv… ▽ More We construct the $Φ^4_3$ measure on an arbitrary 3-dimensional compact Riemannian manifold without boundary as an invariant probability measure of a singular stochastic partial differential equation. Proving the nontriviality and the covariance under Riemannian isometries of that measure gives for the first time a non-perturbative, non-topological interacting Euclidean quantum field theory on curved spaces in dimension 3. This answers a longstanding open problem of constructive quantum field theory on curved 3 dimensional backgrounds. To control analytically several Feynman diagrams appearing in the construction of a number of random fields, we introduce a novel approach of renormalization using microlocal and harmonic analysis. This allows to obtain a renormalized equation which involves some universal constants independent of the manifold. We also define a new vectorial Cole-Hopf transform which allows to deal with the vectorial $Φ^4_3$ model where $Φ$ is now a bundle valued random field. In a companion paper, we develop in a self-contained way all the tools from paradifferential and microlocal analysis that we use to build in our manifold setting a number of analytic and probabilistic objects. △ Less

Submitted 26 June, 2024; v1 submitted 20 April, 2023; originally announced April 2023.

Comments: Section 4 added, Sections 3-4-5 revised to correct a mistake in the previous version

TextANIMAR: Text-based 3D Animal Fine-Grained Retrieval

Authors: Trung-Nghia Le, Tam V. Nguyen, Minh-Quan Le, Trong-Thuan Nguyen, Viet-Tham Huynh, Trong-Le Do, Khanh-Duy Le, Mai-Khiem Tran, Nhat Hoang-Xuan, Thang-Long Nguyen-Ho, Vinh-Tiep Nguyen, Tuong-Nghiem Diep, Khanh-Duy Ho, Xuan-Hieu Nguyen, Thien-Phuc Tran, Tuan-Anh Yang, Kim-Phat Tran, Nhu-Vinh Hoang, Minh-Quang Nguyen, E-Ro Nguyen, Minh-Khoi Nguyen-Nhat, Tuan-An To, Trung-Truc Huynh-Le, Nham-Tan Nguyen, Hoang-Chau Luong , et al. (8 additional authors not shown)

Abstract: 3D object retrieval is an important yet challenging task that has drawn more and more attention in recent years. While existing approaches have made strides in addressing this issue, they are often limited to restricted settings such as image and sketch queries, which are often unfriendly interactions for common users. In order to overcome these limitations, this paper presents a novel SHREC chall… ▽ More 3D object retrieval is an important yet challenging task that has drawn more and more attention in recent years. While existing approaches have made strides in addressing this issue, they are often limited to restricted settings such as image and sketch queries, which are often unfriendly interactions for common users. In order to overcome these limitations, this paper presents a novel SHREC challenge track focusing on text-based fine-grained retrieval of 3D animal models. Unlike previous SHREC challenge tracks, the proposed task is considerably more challenging, requiring participants to develop innovative approaches to tackle the problem of text-based retrieval. Despite the increased difficulty, we believe this task can potentially drive useful applications in practice and facilitate more intuitive interactions with 3D objects. Five groups participated in our competition, submitting a total of 114 runs. While the results obtained in our competition are satisfactory, we note that the challenges presented by this task are far from fully solved. As such, we provide insights into potential areas for future research and improvements. We believe we can help push the boundaries of 3D object retrieval and facilitate more user-friendly interactions via vision-language technologies. https://aichallenge.hcmus.edu.vn/textanimar △ Less

Submitted 9 August, 2023; v1 submitted 12 April, 2023; originally announced April 2023.

Comments: Accepted to Computers and Graphics (3DOR, Journal Track)

SHREC 2022: Fitting and recognition of simple geometric primitives on point clouds

Authors: Chiara Romanengo, Andrea Raffo, Silvia Biasotti, Bianca Falcidieno, Vlassis Fotis, Ioannis Romanelis, Eleftheria Psatha, Konstantinos Moustakas, Ivan Sipiran, Quang-Thuc Nguyen, Chi-Bien Chu, Khoi-Nguyen Nguyen-Ngoc, Dinh-Khoi Vo, Tuan-An To, Nham-Tan Nguyen, Nhat-Quynh Le-Pham, Hai-Dang Nguyen, Minh-Triet Tran, Yifan Qie, Nabil Anwer

Abstract: This paper presents the methods that have participated in the SHREC 2022 track on the fitting and recognition of simple geometric primitives on point clouds. As simple primitives we mean the classical surface primitives derived from constructive solid geometry, i.e., planes, spheres, cylinders, cones and tori. The aim of the track is to evaluate the quality of automatic algorithms for fitting and… ▽ More This paper presents the methods that have participated in the SHREC 2022 track on the fitting and recognition of simple geometric primitives on point clouds. As simple primitives we mean the classical surface primitives derived from constructive solid geometry, i.e., planes, spheres, cylinders, cones and tori. The aim of the track is to evaluate the quality of automatic algorithms for fitting and recognising geometric primitives on point clouds. Specifically, the goal is to identify, for each point cloud, its primitive type and some geometric descriptors. For this purpose, we created a synthetic dataset, divided into a training set and a test set, containing segments perturbed with different kinds of point cloud artifacts. Among the six participants to this track, two are based on direct methods, while four are either fully based on deep learning or combine direct and neural approaches. The performance of the methods is evaluated using various classification and approximation measures. △ Less

Submitted 7 July, 2022; v1 submitted 15 June, 2022; originally announced June 2022.

MSC Class: 68U05; 68U07; 65D18; 65D17 ACM Class: G.1.2; I.3.5; I.5.4

Journal ref: Computers & Graphics 107 (2022) 32-49

6-DoF Pose Estimation of Household Objects for Robotic Manipulation: An Accessible Dataset and Benchmark

Authors: Stephen Tyree, Jonathan Tremblay, Thang To, Jia Cheng, Terry Mosier, Jeffrey Smith, Stan Birchfield

Abstract: We present a new dataset for 6-DoF pose estimation of known objects, with a focus on robotic manipulation research. We propose a set of toy grocery objects, whose physical instantiations are readily available for purchase and are appropriately sized for robotic gras** and manipulation. We provide 3D scanned textured models of these objects, suitable for generating synthetic training data, as wel… ▽ More We present a new dataset for 6-DoF pose estimation of known objects, with a focus on robotic manipulation research. We propose a set of toy grocery objects, whose physical instantiations are readily available for purchase and are appropriately sized for robotic gras** and manipulation. We provide 3D scanned textured models of these objects, suitable for generating synthetic training data, as well as RGBD images of the objects in challenging, cluttered scenes exhibiting partial occlusion, extreme lighting variations, multiple instances per image, and a large variety of poses. Using semi-automated RGBD-to-model texture correspondences, the images are annotated with ground truth poses accurate within a few millimeters. We also propose a new pose evaluation metric called ADD-H based on the Hungarian assignment algorithm that is robust to symmetries in object geometry without requiring their explicit enumeration. We share pre-trained pose estimators for all the toy grocery objects, along with their baseline performance on both validation and test sets. We offer this dataset to the community to help connect the efforts of computer vision researchers with the needs of roboticists. △ Less

Submitted 15 December, 2022; v1 submitted 10 March, 2022; originally announced March 2022.

Comments: IROS 2022. Project page is at https://github.com/swtyree/hope-dataset

Degenerate J-flow on compact Kähler manifolds

Authors: Tat Dat Tô

Abstract: In this note, we study a degenerate twisted J-flow on compact Kähler manifolds. We show that it exists for all time, it is unique and converges to a weak solution of a degenerate twisted J-equation. In particular, this confirms an expectation formulated by Song-Weinkove for the J-flow. As a consequence, we establish the properness of the Mabuchi K-energy twisted by a certain semi-positive closed (… ▽ More In this note, we study a degenerate twisted J-flow on compact Kähler manifolds. We show that it exists for all time, it is unique and converges to a weak solution of a degenerate twisted J-equation. In particular, this confirms an expectation formulated by Song-Weinkove for the J-flow. As a consequence, we establish the properness of the Mabuchi K-energy twisted by a certain semi-positive closed (1,1)-form for Kähler classes in a certain subcone. △ Less

Submitted 3 March, 2023; v1 submitted 23 January, 2022; originally announced January 2022.

Comments: Lemma 3.7 corrected, final version, to appear in Math. Z

Giant and bidirectionally tunable thermopower in non-aqueous ionogels enabled by selective ion do**

Authors: Si**g Liu, Yuewang Yang, He Huang, Jiongzhi Zheng, Gongze Liu, Tsz Ho To, Baoling Huang

Abstract: Ionic thermoelectrics show great potential in low-grade heat harvesting and thermal sensing owing to their ultrahigh thermopower, low cost and ease in production. However, the lack of effective n-type ionic thermoelectric materials seriously hinders their applications. Here, we report giant and bidirectionally tunable thermopowers within an ultrawide range from -23 to +32 mV K-1 at 90% RH in solid… ▽ More Ionic thermoelectrics show great potential in low-grade heat harvesting and thermal sensing owing to their ultrahigh thermopower, low cost and ease in production. However, the lack of effective n-type ionic thermoelectric materials seriously hinders their applications. Here, we report giant and bidirectionally tunable thermopowers within an ultrawide range from -23 to +32 mV K-1 at 90% RH in solid ionic-liquid-based ionogels, rendering it among the best n- and p-type ionic thermoelectric materials. A novel thermopower regulation strategy through ion do** to selectively induce ion aggregates via strong ion-ion interactions is proposed. These charged aggregates are found decisive in modulating the sign and enlarging the magnitude of the thermopower in the ionogels. A prototype wearable device integrated with 12 p-n pairs is demonstrated with a total thermopower of 0.358 V K-1 in general indoor conditions, showing promise for ultrasensitive body heat detection. △ Less

Submitted 24 August, 2021; originally announced August 2021.

Monge-Ampère equations on compact Hessian manifolds

Authors: Vincent Guedj, Tat Dat Tô

Abstract: We consider degenerate Monge-Ampère equations on compact Hessian manifolds. We establish compactness properties of the set of normalized quasi-convex functions and show local and global comparison principles for twisted Monge-Ampère operators. We then use the Perron method to solve Monge-Ampère equations whose RHS involves an arbitrary probability measure, generalizing works of Cheng-Yau, Delanoë,… ▽ More We consider degenerate Monge-Ampère equations on compact Hessian manifolds. We establish compactness properties of the set of normalized quasi-convex functions and show local and global comparison principles for twisted Monge-Ampère operators. We then use the Perron method to solve Monge-Ampère equations whose RHS involves an arbitrary probability measure, generalizing works of Cheng-Yau, Delanoë, Caffarelli-Viaclovsky and Hultgren-Önnheim. The intrinsic approach we develop should be useful in deriving similar results on mildly singular Hessian varieties, in line with the Strominger-Yau-Zaslow conjecture. △ Less

Submitted 28 June, 2021; originally announced June 2021.

Unified modelling of epidemics by coupled dynamics via Monte-Carlo Markov Chain algorithms

Authors: Frédéric Protin, Martel Jules, Duc Thang Nguyen, Hang T. T. Nguyen, Charles Piffault, Willy Rodríguez, Susely Figueroa Iglesias, Tat Dat Tô, Wilderich Tuschmann, Hông Vân Lê, Tenan Yeo, Tien Zung Nguyen

Abstract: To forecast the time dynamics of an epidemic, we propose a discrete stochastic model that unifies and generalizes previous approaches to the subject. Viewing a given population of individuals or groups of individuals with given health state attributes as living in and moving between the nodes of a graph, we use Monte-Carlo Markov Chain techniques to simulate the movements and health state changes… ▽ More To forecast the time dynamics of an epidemic, we propose a discrete stochastic model that unifies and generalizes previous approaches to the subject. Viewing a given population of individuals or groups of individuals with given health state attributes as living in and moving between the nodes of a graph, we use Monte-Carlo Markov Chain techniques to simulate the movements and health state changes of the individuals according to given probabilities of stay that have been preassigned to each of the nodes. We utilize this model to either capture and predict the future geographic evolution of an epidemic in time, or the evolution of an epidemic inside a heterogeneous population which is divided into homogeneous sub-populations, or, more generally, its evolution in a combination or superposition of the previous two contexts. We also prove that when the size of the population increases and a natural hypothesis is satisfied, the stochastic process associated to our model converges to a deterministic process. Indeed, when the length of the time step used in the discrete model converges to zero, in the limit this deterministic process is driven by a differential equation yielding the evolution of the expectation value of the number of infected as a function of time. In the second part of the paper, we apply our model to study the evolution of the Covid-19 epidemic. We deduce a decomposition of the function yielding the number of infectious individuals into "wavelets", which allows to trace in time the expectation value for the number of infections inside each sub-population. Within this framework, we also discuss possible causes for the occurrence of multiple epidemiological waves. △ Less

Submitted 25 June, 2021; originally announced June 2021.

Roughness and Correlations in the Transition from Island to Film Growth: Simulations and Application to CdTe Deposition

Authors: Tung B. T. To, Renan A. L. Almeida, Sukarno O. Ferreira, Fábio D. A. Aarão Reis

Abstract: Using kinetic Monte Carlo simulations, we develop a framework to relate morphological properties and microscopic dynamics during island growth, coalescence, and initial formation of continuous heteroepitaxial films. The average island width is controlled by adatom mobility on the substrate. Subsequent evolution strongly depends on the Ehrlich-Schwöebel energy barrier $E_{ES}$ of the deposited mate… ▽ More Using kinetic Monte Carlo simulations, we develop a framework to relate morphological properties and microscopic dynamics during island growth, coalescence, and initial formation of continuous heteroepitaxial films. The average island width is controlled by adatom mobility on the substrate. Subsequent evolution strongly depends on the Ehrlich-Schwöebel energy barrier $E_{ES}$ of the deposited material. As $E_{ES}$ decreases, islands becomes taller and their coalescence is delayed. For small islands and large $E_{ES}$, the global roughness increases as $W\sim{\text{thickness}}^{1/2}$ and the local roughness increases at short scales (apparent anomalous scaling) before and after island coalescence. If the islands are wide, $W$ may have a plateau for large $E_{ES}$ and has a maximum for small $E_{ES}$ when the islands coalesce. This framework is applied to atomic-force microscopy data of the initial stages of CdTe deposition on Kapton: a maximum of $W$ during island coalescence indicates negligible ES barrier, consistently with scaling properties of much thicker films, and the diffusion coefficient ${10}^{-7}{\text{--}}{10}^{-5}{\text{cm}}^2/{\text s}$ on the Kapton surface at $150\,^{\circ}\mathrm{C}$ is estimated. Applications of the framework to other materials are suggested, in which the expected roles of ES barriers are highlighted. △ Less

Submitted 21 July, 2021; v1 submitted 21 May, 2021; originally announced May 2021.

Journal ref: Applied Surface Science 560 (2021) 149946

Shear dynamics of confined membranes

Authors: Thomas Le Goff, Tung B. T. To, Olivier Pierre-Louis

Abstract: We model the nonlinear response of a lubricated contact composed of a two-dimensional lipid membrane immersed in a simple fluid between two parallel flat and porous walls under shear. The nonlinear dynamics of the membrane gives rise to a rich dynamical behavior depending on the shear velocity. In quiescent conditions (i.e., absence of shear), the membrane freezes into a disordered labyrinthine wr… ▽ More We model the nonlinear response of a lubricated contact composed of a two-dimensional lipid membrane immersed in a simple fluid between two parallel flat and porous walls under shear. The nonlinear dynamics of the membrane gives rise to a rich dynamical behavior depending on the shear velocity. In quiescent conditions (i.e., absence of shear), the membrane freezes into a disordered labyrinthine wrinkle pattern. In the presence of shear, we find four different regimes depending on the shear rate. Regime I. For small shear, the labyrinthine pattern is still frozen, but exhibits a small drift which is mainly along the shear direction. In this regime, the tangential forces on the walls due to the presence of the membrane increase linearly with the shear rate. Regime II. When the shear rate is increased above a critical value, the membrane rearranges, and wrinkles start to align along the shear direction. This regime is accompanied by a sharp drop of the tangential forces on the wall. Regime III. For larger shear rates, the wrinkles align strongly along the shear direction, with a set of dislocation defects which assemble in pairs. The tangential forces are then controlled by the number of dislocations, and by the number of wrinkles between the two dislocations within each dislocation pairs. In this dislocation-dominated regime, the tangential forces in the transverse direction most often exceed those in the shear direction. Regime IV. For even larger shear, the membrane organizes into a perfect array of parallel stripes with no defects. The wavelength of the wrinkles is still identical to the wavelength in the absence of shear. In this final regime, the tangential forces due to the membrane vanish. These behaviors give rise to a non-linear rheological behavior of lubricated contacts containing membranes. △ Less

Submitted 21 May, 2021; originally announced May 2021.

Comments: Soft Matter (2021)

Journal ref: Soft Matter 17 (2021) 5467

Molecular field theory for biaxial smectic A liquid crystals

Authors: Tung B. T. To, Timothy J. Sluckin, Geoffrey R. Luckhurst

Abstract: Thermotropic biaxial nematic phases seem to be rare, but biaxial smectic A phases less so. Here we use molecular field theory to study a simple two-parameter model, with one parameter promoting a biaxial phase and the second promoting smecticity. The theory combines the biaxial Maier-Saupe and McMillan models. We use alternatively the Sonnet-Virga-Durand (SVD) and geometric mean approximations (GM… ▽ More Thermotropic biaxial nematic phases seem to be rare, but biaxial smectic A phases less so. Here we use molecular field theory to study a simple two-parameter model, with one parameter promoting a biaxial phase and the second promoting smecticity. The theory combines the biaxial Maier-Saupe and McMillan models. We use alternatively the Sonnet-Virga-Durand (SVD) and geometric mean approximations (GMA) to characterize molecular biaxiality by a single parameter. For non-zero smecticity and biaxiality, the model always predicts a ground state biaxial smectic A phase. For a low degree of smectic order, the phase diagram is very rich, predicting uniaxial and biaxial nematic and smectic phases, with in addition a variety of tricritical and tetracritical points. For higher degrees of smecticity, the region of stability of the biaxial nematic phase is restricted and eventually disappears, yielding to the biaxial smectic phase. Phase diagrams from the two alternative approximations for molecular biaxiality are similar, except inasmuch that SVD allows for a first order isotropic-nematic biaxial transition, whereas GMA predicts a Landau point separating isotropic and biaxial nematic phases. We speculate that the rarity of thermotropic biaxial nematic phases is partly a consequence of the presence of stabler analogous smectic phases. △ Less

Submitted 20 May, 2021; originally announced May 2021.

Journal ref: The Journal of Chemical Physics, 139, 134902 (2013)

A fast and simple modification of Newton's method hel** to avoid saddle points

Authors: Tuyen Trung Truong, Tat Dat To, Tuan Hang Nguyen, Thu Hang Nguyen, Hoang Phuong Nguyen, Maged Helmy

Abstract: We propose in this paper New Q-Newton's method. The update rule is very simple conceptually, for example $x_{n+1}=x_n-w_n$ where $w_n=pr_{A_n,+}(v_n)-pr_{A_n,-}(v_n)$, with $A_n=\nabla ^2f(x_n)+δ_n||\nabla f(x_n)||^2.Id$ and $v_n=A_n^{-1}.\nabla f(x_n)$. Here $δ_n$ is an appropriate real number so that $A_n$ is invertible, and $pr_{A_n,\pm}$ are projections to the vector subspaces generated by eig… ▽ More We propose in this paper New Q-Newton's method. The update rule is very simple conceptually, for example $x_{n+1}=x_n-w_n$ where $w_n=pr_{A_n,+}(v_n)-pr_{A_n,-}(v_n)$, with $A_n=\nabla ^2f(x_n)+δ_n||\nabla f(x_n)||^2.Id$ and $v_n=A_n^{-1}.\nabla f(x_n)$. Here $δ_n$ is an appropriate real number so that $A_n$ is invertible, and $pr_{A_n,\pm}$ are projections to the vector subspaces generated by eigenvectors of positive (correspondingly negative) eigenvalues of $A_n$. The main result of this paper roughly says that if $f$ is $C^3$ (can be unbounded from below) and a sequence $\{x_n\}$, constructed by the New Q-Newton's method from a random initial point $x_0$, {\bf converges}, then the limit point is a critical point and is not a saddle point, and the convergence rate is the same as that of Newton's method. The first author has recently been successful incorporating Backtracking line search to New Q-Newton's method, thus resolving the convergence guarantee issue observed for some (non-smooth) cost functions. An application to quickly finding zeros of a univariate meromorphic function will be discussed. Various experiments are performed, against well known algorithms such as BFGS and Adaptive Cubic Regularization are presented. △ Less

Submitted 9 September, 2021; v1 submitted 2 June, 2020; originally announced June 2020.

Comments: Main part: 43 pages, Appendix: 21 pages. Some misprints corrected. Add comparison to a relevant circle of ideas by Goldfarb, Gould et al, and others. Add applications to finding roots of a univariate meromorphic function

Thin film growth models with long surface diffusion lengths

Authors: Tung B. T. To, Vitor B. de Sousa, Fábio D. A. Aarão Reis

Abstract: In limited mobility (LM) models of thin film deposition, the final position of each atom or molecule is chosen according to a set of stochastic rules before the incidence of another atom or molecule. Here we investigate the possibility of a LM model to reproduce features of a more realistic approach that represents the interplay of collective adatom diffusion and the external flux. In the LM model… ▽ More In limited mobility (LM) models of thin film deposition, the final position of each atom or molecule is chosen according to a set of stochastic rules before the incidence of another atom or molecule. Here we investigate the possibility of a LM model to reproduce features of a more realistic approach that represents the interplay of collective adatom diffusion and the external flux. In the LM model introduced here, each adatom may execute $G$ hops to neighboring columns of the deposit, but a hop attempt from a site with $n$ lateral neighbors has probability $P^n$, with $P<1$. These rules resemble those of the Clarke-Vvedensky (CV) model without energy barriers at step edges, whose main parameters are the diffusion-to-deposition ratio $R$ on terraces and the detachment probability $ε$ per lateral neighbor. At short times, the roughness of the LM model can be written in terms of a scaling function of $G$ and $P$ and the growth exponent is consistent with the Villain-Lai-Das Sarma universality class. The evolution of the surface roughness and of the autocorrelation function of the CV model is reproduced with reasonable accuracy by the LM model with suitable choices of parameters. The increase of the parameters $G$ and $R$ of those models produces smoother film surfaces, while the increase of $P$ and $ε$ smoothen the terrace boundaries at short lengthscales. However, the detachment probabilities of the two models have very different effects on the surface roughness: in the LM model, for fixed $G$, the surface roughness increases as $P$ increases; in the CV model, the surface smoothens as $ε$ increases, for fixed $R$. This result is related to the non-Markovian nature of the LM model, since the diffusivity of an adatom depends on its history at the film surface and may be severely reduced after a detachment from a terrace step. △ Less

Submitted 8 May, 2020; v1 submitted 9 April, 2020; originally announced April 2020.

Comments: Preprint for a publication in Physica A

Journal ref: Physica A 511 (2018) 240-250

Domain formation in the deposition of thin films of two-component mixtures

Authors: Tung B. T. To, Fábio D. A. Aarão Reis

Abstract: We perform a kinetic Monte Carlo simulation study of a model of thin film deposition of a two-component mixture in which the activation energy for diffusion of an adatom is additive over its nearest neighbors and in which the interactions between adatoms of the same species are stronger than those between adatoms of different species. The film morphology is investigated for broad ranges of values… ▽ More We perform a kinetic Monte Carlo simulation study of a model of thin film deposition of a two-component mixture in which the activation energy for diffusion of an adatom is additive over its nearest neighbors and in which the interactions between adatoms of the same species are stronger than those between adatoms of different species. The film morphology is investigated for broad ranges of values of the ratios between terrace diffusivity and atomic flux, of the probabilities of detachment from lateral neighbors, and of probabilities of crossing terrace edges. First, we consider a symmetric case in which the interactions of adatoms of the same species are the same for the two components. For low and intermediate temperatures, we show the formation of narrow domains, whose widths are $3$-$30$ lattice constants, which meander in the layers parallel to the substrate, and are connected through very long distances. The domain width depends on the diffusivity-flux ratio of the same species, but it is weakly affected by interactions of different species, so rough estimates of that width can be obtained by using tabulated properties of films with a single component. At high temperatures, the separation of domains is enhanced, but their long distance connectivity is lost, which shows that the formation of the long meandering domains is restricted to a certain range of temperature and flux. Similar morphological features are obtained when the intra-species interactions are different and the adatom diffusion coefficients on terraces of the same species differ up to two orders of magnitude. An approximate scaling relation for the domain width is obtained, but the species with the largest mobility constrains the temperature range in which the long connected domains are found. △ Less

Submitted 8 May, 2020; v1 submitted 9 April, 2020; originally announced April 2020.

Comments: Preprint accepted for publication by Journal of Alloys and Compounds

Journal ref: Journal of Alloys and Compounds 835 (2020) 155093(1-14)

Stability and Hölder regularity of solutions to complex Monge-Ampère equations on compact Hermitian manifolds

Authors: Chinh H. Lu, Trong-Thuc Phung, Tât-Dat Tô

Abstract: Let $(X,ω)$ be a compact Hermitian manifold. We establish a stability result for solutions to complex Monge-Ampère equations with right-hand side in $L^p$, $p>1$. Using this we prove that the solutions are Hölder continuous with the same exponent as in the Kähler case \cite{DDGKPZ14}. Our techniques also apply to the setting of big cohomology classes on compact Kähler manifolds. Let $(X,ω)$ be a compact Hermitian manifold. We establish a stability result for solutions to complex Monge-Ampère equations with right-hand side in $L^p$, $p>1$. Using this we prove that the solutions are Hölder continuous with the same exponent as in the Kähler case \cite{DDGKPZ14}. Our techniques also apply to the setting of big cohomology classes on compact Kähler manifolds. △ Less

Submitted 14 November, 2020; v1 submitted 18 March, 2020; originally announced March 2020.

Comments: final version to appear in Annales de l'Institut Fourier

MSC Class: 32W20; 32U05; 32Q15

Convergence of the Hesse-Koszul flow on compact Hessian manifolds

Authors: Stéphane Puechmorel, Tat Dat Tô

Abstract: We study the long time behavior of the Hesse-Koszul flow on compact Hessian manifolds. When the first affine Chern class is negative, we prove that the flow converges to the unique Hesse-Einstein metric. We also derive a convergence result for a twisted Hesse-Koszul flow on any compact Hessian manifold. These results give alternative proofs for the existence of the unique Hesse-Einstein metric by… ▽ More We study the long time behavior of the Hesse-Koszul flow on compact Hessian manifolds. When the first affine Chern class is negative, we prove that the flow converges to the unique Hesse-Einstein metric. We also derive a convergence result for a twisted Hesse-Koszul flow on any compact Hessian manifold. These results give alternative proofs for the existence of the unique Hesse-Einstein metric by Cheng-Yau and Caffarelli-Viaclovsky as well as the real Calabi theorem by Cheng-Yau, Delanoë and Caffarelli-Viaclovsky. △ Less

Submitted 9 January, 2020; originally announced January 2020.

Comments: 26 pages

Camera-to-Robot Pose Estimation from a Single Image

Authors: Timothy E. Lee, Jonathan Tremblay, Thang To, Jia Cheng, Terry Mosier, Oliver Kroemer, Dieter Fox, Stan Birchfield

Abstract: We present an approach for estimating the pose of an external camera with respect to a robot using a single RGB image of the robot. The image is processed by a deep neural network to detect 2D projections of keypoints (such as joints) associated with the robot. The network is trained entirely on simulated data using domain randomization to bridge the reality gap. Perspective-n-point (PnP) is then… ▽ More We present an approach for estimating the pose of an external camera with respect to a robot using a single RGB image of the robot. The image is processed by a deep neural network to detect 2D projections of keypoints (such as joints) associated with the robot. The network is trained entirely on simulated data using domain randomization to bridge the reality gap. Perspective-n-point (PnP) is then used to recover the camera extrinsics, assuming that the camera intrinsics and joint configuration of the robot manipulator are known. Unlike classic hand-eye calibration systems, our method does not require an off-line calibration step. Rather, it is capable of computing the camera extrinsics from a single frame, thus opening the possibility of on-line calibration. We show experimental results for three different robots and camera sensors, demonstrating that our approach is able to achieve accuracy with a single frame that is comparable to that of classic off-line hand-eye calibration using multiple frames. With additional frames from a static pose, accuracy improves even further. Code, datasets, and pretrained models for three widely-used robot manipulators are made available. △ Less

Submitted 23 April, 2020; v1 submitted 20 November, 2019; originally announced November 2019.

Comments: ICRA 2020. Project page is at https://research.nvidia.com/publication/2020-03_DREAM

Toward Sim-to-Real Directional Semantic Gras**

Authors: Shariq Iqbal, Jonathan Tremblay, Thang To, Jia Cheng, Erik Leitch, Andy Campbell, Kirby Leung, Duncan McKay, Stan Birchfield

Abstract: We address the problem of directional semantic gras**, that is, gras** a specific object from a specific direction. We approach the problem using deep reinforcement learning via a double deep Q-network (DDQN) that learns to map downsampled RGB input images from a wrist-mounted camera to Q-values, which are then translated into Cartesian robot control commands via the cross-entropy method (CEM)… ▽ More We address the problem of directional semantic gras**, that is, gras** a specific object from a specific direction. We approach the problem using deep reinforcement learning via a double deep Q-network (DDQN) that learns to map downsampled RGB input images from a wrist-mounted camera to Q-values, which are then translated into Cartesian robot control commands via the cross-entropy method (CEM). The network is learned entirely on simulated data generated by a custom robot simulator that models both physical reality (contacts) and perceptual quality (high-quality rendering). The reality gap is bridged using domain randomization. The system is an example of end-to-end (map** input monocular RGB images to output Cartesian motor commands) gras** of objects from multiple pre-defined object-centric orientations, such as from the side or top. We show promising results in both simulation and the real world, along with some challenges faced and the need for future research in this area. △ Less

Submitted 18 August, 2020; v1 submitted 4 September, 2019; originally announced September 2019.

Comments: ICRA 2020. Video is at https://youtu.be/bjJLtNdVj9w

Viscosity solutions to parabolic complex Monge-Ampère equations

Authors: Hoang-Son Do, Giang Le, Tat Dat Tô

Abstract: In this paper, we study the Cauchy-Dirichlet problem for Parabolic complex Monge-Ampère equations on a strongly pseudoconvex domain by the viscosity method. We extend the results in [EGZ15b] on the existence of solution and the convergence at infinity. We also establish the Hölder regularity of the solutions when the Cauchy-Dirichlet data are Hölder continuous. In this paper, we study the Cauchy-Dirichlet problem for Parabolic complex Monge-Ampère equations on a strongly pseudoconvex domain by the viscosity method. We extend the results in [EGZ15b] on the existence of solution and the convergence at infinity. We also establish the Hölder regularity of the solutions when the Cauchy-Dirichlet data are Hölder continuous. △ Less

Submitted 25 November, 2019; v1 submitted 25 May, 2019; originally announced May 2019.

Comments: 35 pages. arXiv admin note: text overlap with arXiv:1407.2494 by other authors

Convergence of the weak Kähler-Ricci Flow on manifolds of general type

Authors: Tat Dat Tô

Abstract: We study the Kähler-Ricci flow on compact Kähler manifolds whose canonical bundle is big. We show that the normalized Kähler-Ricci flow has long time existence in the viscosity sense, is continuous in a Zariski open set, and converges to the unique singular Kähler-Einstein metric in the canonical class. The key ingredient is a viscosity theory for degenerate complex Monge-Ampère flows in big class… ▽ More We study the Kähler-Ricci flow on compact Kähler manifolds whose canonical bundle is big. We show that the normalized Kähler-Ricci flow has long time existence in the viscosity sense, is continuous in a Zariski open set, and converges to the unique singular Kähler-Einstein metric in the canonical class. The key ingredient is a viscosity theory for degenerate complex Monge-Ampère flows in big classes that we develop, extending and refining the approach of Eyssidieux-Guedj-Zeriahi. △ Less

Submitted 10 November, 2023; v1 submitted 3 May, 2019; originally announced May 2019.

Comments: details added in proof of Theorem 4.3

Journal ref: Int. Math. Res. Not. 2021 No. 8, 6373-6404

Deep Object Pose Estimation for Semantic Robotic Gras** of Household Objects

Authors: Jonathan Tremblay, Thang To, Balakumar Sundaralingam, Yu Xiang, Dieter Fox, Stan Birchfield

Abstract: Using synthetic data for training deep neural networks for robotic manipulation holds the promise of an almost unlimited amount of pre-labeled training data, generated safely out of harm's way. One of the key challenges of synthetic data, to date, has been to bridge the so-called reality gap, so that networks trained on synthetic data operate correctly when exposed to real-world data. We explore t… ▽ More Using synthetic data for training deep neural networks for robotic manipulation holds the promise of an almost unlimited amount of pre-labeled training data, generated safely out of harm's way. One of the key challenges of synthetic data, to date, has been to bridge the so-called reality gap, so that networks trained on synthetic data operate correctly when exposed to real-world data. We explore the reality gap in the context of 6-DoF pose estimation of known objects from a single RGB image. We show that for this problem the reality gap can be successfully spanned by a simple combination of domain randomized and photorealistic data. Using synthetic data generated in this manner, we introduce a one-shot deep neural network that is able to perform competitively against a state-of-the-art network trained on a combination of real and synthetic data. To our knowledge, this is the first deep network trained only on synthetic data that is able to achieve state-of-the-art performance on 6-DoF object pose estimation. Our network also generalizes better to novel environments including extreme lighting conditions, for which we show qualitative results. Using this network we demonstrate a real-time system estimating object poses with sufficient accuracy for real-world semantic gras** of known household objects in clutter by a real robot. △ Less

Submitted 27 September, 2018; originally announced September 2018.

Comments: Conference on Robot Learning (CoRL) 2018

Adhesion dynamics of confined membranes

Authors: Tung B. T. To, Thomas Le Goff, Olivier Pierre-Louis

Abstract: We report on the modeling of the dynamics of confined lipid membranes. We derive a thin film model in the lubrication limit which describes an inextensible liquid membrane with bending rigidity confined between two adhesive walls. The resulting equations share similarities with the Swift-Hohenberg model. However, inextensibility is enforced by a time-dependent nonlocal tension. Depending on the ex… ▽ More We report on the modeling of the dynamics of confined lipid membranes. We derive a thin film model in the lubrication limit which describes an inextensible liquid membrane with bending rigidity confined between two adhesive walls. The resulting equations share similarities with the Swift-Hohenberg model. However, inextensibility is enforced by a time-dependent nonlocal tension. Depending on the excess membrane area available in the system, three different dynamical regimes, denoted as A, B and C, are found from the numerical solution of the model. In regime A, membranes with small excess area form flat adhesion domains and freeze. Such freezing is interpreted by means of an effective model for curvature-driven domain wall motion. The nonlocal membrane tension tends to a negative value corresponding to the linear stability threshold of flat domain walls in the Swift-Hohenberg equation. In regime B, membranes with intermediate excess areas exhibit endless coarsening with coexistence of flat adhesion domains and wrinkle domains. The tension tends to the nonlinear stability threshold of flat domain walls in the Swift-Hohenberg equation. The fraction of the system covered by the wrinkle phase increases linearly with the excess area in regime B. In regime C, membranes with large excess area are completely covered by a frozen labyrinthine pattern of wrinkles. As the excess area is increased, the tension increases and the wavelength of the wrinkles decreases. For large membrane area, there is a crossover to a regime where the extrema of the wrinkles are in contact with the walls. In all regimes after an initial transient, robust localised structures form, leading to an exact conservation of the number of adhesion domains. △ Less

Submitted 1 August, 2018; originally announced August 2018.

Synthetically Trained Neural Networks for Learning Human-Readable Plans from Real-World Demonstrations

Authors: Jonathan Tremblay, Thang To, Artem Molchanov, Stephen Tyree, Jan Kautz, Stan Birchfield

Abstract: We present a system to infer and execute a human-readable program from a real-world demonstration. The system consists of a series of neural networks to perform perception, program generation, and program execution. Leveraging convolutional pose machines, the perception network reliably detects the bounding cuboids of objects in real images even when severely occluded, after training only on synth… ▽ More We present a system to infer and execute a human-readable program from a real-world demonstration. The system consists of a series of neural networks to perform perception, program generation, and program execution. Leveraging convolutional pose machines, the perception network reliably detects the bounding cuboids of objects in real images even when severely occluded, after training only on synthetic images using domain randomization. To increase the applicability of the perception network to new scenarios, the network is formulated to predict in image space rather than in world space. Additional networks detect relationships between objects, generate plans, and determine actions to reproduce a real-world demonstration. The networks are trained entirely in simulation, and the system is tested in the real world on the pick-and-place problem of stacking colored cubes using a Baxter robot. △ Less

Submitted 10 July, 2018; v1 submitted 18 May, 2018; originally announced May 2018.

Comments: IEEE International Conference on Robotics and Automation (ICRA) 2018. For associated video, see https://youtu.be/B7ZT5oSnRys

Falling Things: A Synthetic Dataset for 3D Object Detection and Pose Estimation

Authors: Jonathan Tremblay, Thang To, Stan Birchfield

Abstract: We present a new dataset, called Falling Things (FAT), for advancing the state-of-the-art in object detection and 3D pose estimation in the context of robotics. By synthetically combining object models and backgrounds of complex composition and high graphical quality, we are able to generate photorealistic images with accurate 3D pose annotations for all objects in all images. Our dataset contains… ▽ More We present a new dataset, called Falling Things (FAT), for advancing the state-of-the-art in object detection and 3D pose estimation in the context of robotics. By synthetically combining object models and backgrounds of complex composition and high graphical quality, we are able to generate photorealistic images with accurate 3D pose annotations for all objects in all images. Our dataset contains 60k annotated photos of 21 household objects taken from the YCB dataset. For each image, we provide the 3D poses, per-pixel class segmentation, and 2D/3D bounding box coordinates for all objects. To facilitate testing different input modalities, we provide mono and stereo RGB images, along with registered dense depth images. We describe in detail the generation process and statistical analysis of the data. △ Less

Submitted 10 July, 2018; v1 submitted 17 April, 2018; originally announced April 2018.

Comments: CVPR 2018 Workshop on Real World Challenges and New Benchmarks for Deep Learning in Robotic Vision

Training Deep Networks with Synthetic Data: Bridging the Reality Gap by Domain Randomization

Authors: Jonathan Tremblay, Aayush Prakash, David Acuna, Mark Brophy, Varun Jampani, Cem Anil, Thang To, Eric Cameracci, Shaad Boochoon, Stan Birchfield

Abstract: We present a system for training deep neural networks for object detection using synthetic images. To handle the variability in real-world data, the system relies upon the technique of domain randomization, in which the parameters of the simulator$-$such as lighting, pose, object textures, etc.$-$are randomized in non-realistic ways to force the neural network to learn the essential features of th… ▽ More We present a system for training deep neural networks for object detection using synthetic images. To handle the variability in real-world data, the system relies upon the technique of domain randomization, in which the parameters of the simulator$-$such as lighting, pose, object textures, etc.$-$are randomized in non-realistic ways to force the neural network to learn the essential features of the object of interest. We explore the importance of these parameters, showing that it is possible to produce a network with compelling performance using only non-artistically-generated synthetic data. With additional fine-tuning on real data, the network yields better performance than using real data alone. This result opens up the possibility of using inexpensive synthetic data for training neural networks while avoiding the need to collect large amounts of hand-annotated real-world data or to generate high-fidelity synthetic worlds$-$both of which remain bottlenecks for many applications. The approach is evaluated on bounding box detection of cars on the KITTI dataset. △ Less

Submitted 23 April, 2018; v1 submitted 17 April, 2018; originally announced April 2018.

Comments: CVPR 2018 Workshop on Autonomous Driving

A viscosity approach to the Dirichlet problem for degenerate complex Hessian type equations

Authors: Slawomir Dinew, Hoang-Son Do, Tat Dat To

Abstract: A viscosity approach is introduced for the Dirichlet problem associated to complex Hessian type equations on domains in $\C^n$. The arguments are modelled on the theory of viscosity solutions for real Hessian type equations developed by Trudinger. As consequence we solve the Dirichlet problem for the Hessian quotient and special Lagrangian equations. We also establish basic regularity results for… ▽ More A viscosity approach is introduced for the Dirichlet problem associated to complex Hessian type equations on domains in $\C^n$. The arguments are modelled on the theory of viscosity solutions for real Hessian type equations developed by Trudinger. As consequence we solve the Dirichlet problem for the Hessian quotient and special Lagrangian equations. We also establish basic regularity results for the solutions. △ Less

Submitted 22 December, 2017; originally announced December 2017.

Comments: 29 pages

MSC Class: 32U05; 35J70

Journal ref: Analysis & PDE 12 (2019) 505-535

Fully non-linear parabolic equations on compact Hermitian manifolds

Authors: Duong H. Phong, Dat T. Tô

Abstract: A notion of parabolic C-subsolutions is introduced for parabolic equations, extending the theory of C-subsolutions recently developed by B. Guan and more specifically G. Székelyhidi for elliptic equations. The resulting parabolic theory provides a convenient unified approach for the study of many geometric flows. A notion of parabolic C-subsolutions is introduced for parabolic equations, extending the theory of C-subsolutions recently developed by B. Guan and more specifically G. Székelyhidi for elliptic equations. The resulting parabolic theory provides a convenient unified approach for the study of many geometric flows. △ Less

Submitted 14 December, 2017; v1 submitted 29 November, 2017; originally announced November 2017.

Comments: References added, typos fixed

Demonstration of Universal Parametric Entangling Gates on a Multi-Qubit Lattice

Authors: M. Reagor, C. B. Osborn, N. Tezak, A. Staley, G. Prawiroatmodjo, M. Scheer, N. Alidoust, E. A. Sete, N. Didier, M. P. da Silva, E. Acala, J. Angeles, A. Bestwick, M. Block, B. Bloom, A. Bradley, C. Bui, S. Caldwell, L. Capelluto, R. Chilcott, J. Cordova, G. Crossman, M. Curtis, S. Deshpande, T. El Bouayadi , et al. (34 additional authors not shown)

Abstract: We show that parametric coupling techniques can be used to generate selective entangling interactions for multi-qubit processors. By inducing coherent population exchange between adjacent qubits under frequency modulation, we implement a universal gateset for a linear array of four superconducting qubits. An average process fidelity of $\mathcal{F}=93\%$ is estimated for three two-qubit gates via… ▽ More We show that parametric coupling techniques can be used to generate selective entangling interactions for multi-qubit processors. By inducing coherent population exchange between adjacent qubits under frequency modulation, we implement a universal gateset for a linear array of four superconducting qubits. An average process fidelity of $\mathcal{F}=93\%$ is estimated for three two-qubit gates via quantum process tomography. We establish the suitability of these techniques for computation by preparing a four-qubit maximally entangled state and comparing the estimated state fidelity against the expected performance of the individual entangling gates. In addition, we prepare an eight-qubit register in all possible bitstring permutations and monitor the fidelity of a two-qubit gate across one pair of these qubits. Across all such permutations, an average fidelity of $\mathcal{F}=91.6\pm2.6\%$ is observed. These results thus offer a path to a scalable architecture with high selectivity and low crosstalk. △ Less

Submitted 26 February, 2018; v1 submitted 20 June, 2017; originally announced June 2017.

Journal ref: Science Advances 02 Feb 2018: Vol. 4, no. 2, eaao3603

Parametrically Activated Entangling Gates Using Transmon Qubits

Authors: S. Caldwell, N. Didier, C. A. Ryan, E. A. Sete, A. Hudson, P. Karalekas, R. Manenti, M. Reagor, M. P. da Silva, R. Sinclair, E. Acala, N. Alidoust, J. Angeles, A. Bestwick, M. Block, B. Bloom, A. Bradley, C. Bui, L. Capelluto, R. Chilcott, J. Cordova, G. Crossman, M. Curtis, S. Deshpande, T. El Bouayadi , et al. (37 additional authors not shown)

Abstract: We describe and implement a family of entangling gates activated by radio-frequency flux modulation applied to a tunable transmon that is statically coupled to a neighboring transmon. The effect of this modulation is the resonant exchange of photons directly between levels of the two-transmon system, obviating the need for mediating qubits or resonator modes and allowing for the full utilization o… ▽ More We describe and implement a family of entangling gates activated by radio-frequency flux modulation applied to a tunable transmon that is statically coupled to a neighboring transmon. The effect of this modulation is the resonant exchange of photons directly between levels of the two-transmon system, obviating the need for mediating qubits or resonator modes and allowing for the full utilization of all qubits in a scalable architecture. The resonance condition is selective in both the frequency and amplitude of modulation and thus alleviates frequency crowding. We demonstrate the use of three such resonances to produce entangling gates that enable universal quantum computation: one iSWAP gate and two distinct controlled Z gates. We report interleaved randomized benchmarking results indicating gate error rates of 6% for the iSWAP (duration 135ns) and 9% for the controlled Z gates (durations 175 ns and 270 ns), limited largely by qubit coherence. △ Less

Submitted 8 December, 2017; v1 submitted 20 June, 2017; originally announced June 2017.

Comments: As submitted. 7 pages, 6 figures

Journal ref: Phys. Rev. Applied 10, 034050 (2018)

Regularizing properties of Complex Monge-Ampère flows II: Hermitian manifolds

Authors: Tat Dat Tô

Abstract: We prove that a general complex Monge-Ampère flow on a Hermitian manifold can be run from an arbitrary initial condition with zero Lelong number at all points. Using this property, we confirm a conjecture of Tosatti-Weinkove: the Chern-Ricci flow performs a canonical surgical contraction. Finally, we study a generalization of the Chern-Ricci flow on compact Hermitian manifolds, namely the twisted… ▽ More We prove that a general complex Monge-Ampère flow on a Hermitian manifold can be run from an arbitrary initial condition with zero Lelong number at all points. Using this property, we confirm a conjecture of Tosatti-Weinkove: the Chern-Ricci flow performs a canonical surgical contraction. Finally, we study a generalization of the Chern-Ricci flow on compact Hermitian manifolds, namely the twisted Chern-Ricci flow. △ Less

Submitted 31 July, 2017; v1 submitted 15 January, 2017; originally announced January 2017.

Comments: 32 pages, final version, to appear in Math. Ann

Journal ref: Mathematische Annalen, 372 (2018), Issue 1-2, pp 699-741

Regularizing properties of Complex Monge-Ampère flows

Authors: Tat Dat Tô

Abstract: We study the regularizing properties of complex Monge-Ampère flows on a Kähler manifold $(X,ω)$ when the initial data are $ω$-psh functions with zero Lelong number at all points. We prove that the general Monge-Ampère flow has a solution which is immediately smooth. We also prove the uniqueness and stability of solution. We study the regularizing properties of complex Monge-Ampère flows on a Kähler manifold $(X,ω)$ when the initial data are $ω$-psh functions with zero Lelong number at all points. We prove that the general Monge-Ampère flow has a solution which is immediately smooth. We also prove the uniqueness and stability of solution. △ Less

Submitted 21 April, 2016; originally announced April 2016.

Comments: 30 pages

Journal ref: J Funct Anal 272 (2017), no 5, 2058-2091

Optimal Regularity and the Free Boundary in the Parabolic Signorini Problem

Authors: Donatella Danielli, Nicola Garofalo, Arshak Petrosyan, Tung To

Abstract: We give a comprehensive treatment of the parabolic Signorini problem based on a generalization of Almgren's monotonicity of the frequency. This includes the proof of the optimal regularity of solutions, classification of free boundary points, the regularity of the regular set and the structure of the singular set. We give a comprehensive treatment of the parabolic Signorini problem based on a generalization of Almgren's monotonicity of the frequency. This includes the proof of the optimal regularity of solutions, classification of free boundary points, the regularity of the regular set and the structure of the singular set. △ Less

Submitted 25 June, 2013; v1 submitted 21 June, 2013; originally announced June 2013.

Comments: 91 pages, 4 figures

MSC Class: 35R35; 35K85

On a conjecture of compatibility of multi-states characters

Authors: Michel Habib, Thu-Hien To

Abstract: Perfect phylogeny consisting of determining the compatibility of a set of characters is known to be NP-complete. We propose in this article a conjecture on the necessary and sufficient conditions of compatibility: Given a set $\mathcal{C}$ of $r$-states full characters, there exists a function $f(r)$ such that $\mathcal{C}$ is compatible iff every set of $f(r)$ characters of $\mathcal{C}$ is compa… ▽ More Perfect phylogeny consisting of determining the compatibility of a set of characters is known to be NP-complete. We propose in this article a conjecture on the necessary and sufficient conditions of compatibility: Given a set $\mathcal{C}$ of $r$-states full characters, there exists a function $f(r)$ such that $\mathcal{C}$ is compatible iff every set of $f(r)$ characters of $\mathcal{C}$ is compatible. Some previous work showed that $f(2)=2$, $f(3)=3$ and $f(r) \ge r-1$. Gusfield et al. 09 conjectured that $f(r) = r$ for any $r \ge 2$. In this paper, we present an example showing that $f(4) \ge 5$ and then a closure operation for chordal sandwich graphs. The later problem is a common approach of perfect phylogeny. This operation can be the first step to simplify the problem before solving some particular cases $f(4), f(5), ... $, and determining the function $f(r)$. △ Less

Submitted 5 May, 2011; originally announced May 2011.

Constructing a Minimum-Level Phylogenetic Network from a Dense Triplet Set in Polynomial Time

Authors: Michel Habib, Thu-Hien To

Abstract: For a given set $\mathcal{L}$ of species and a set $\mathcal{T}$ of triplets on $\mathcal{L}$, one wants to construct a phylogenetic network which is consistent with $\mathcal{T}$, i.e which represents all triplets of $\mathcal{T}$. The level of a network is defined as the maximum number of hybrid vertices in its biconnected components. When $\mathcal{T}$ is dense, there exist polynomial time algo… ▽ More For a given set $\mathcal{L}$ of species and a set $\mathcal{T}$ of triplets on $\mathcal{L}$, one wants to construct a phylogenetic network which is consistent with $\mathcal{T}$, i.e which represents all triplets of $\mathcal{T}$. The level of a network is defined as the maximum number of hybrid vertices in its biconnected components. When $\mathcal{T}$ is dense, there exist polynomial time algorithms to construct level-$0,1,2$ networks (Aho et al. 81, Jansson et al. 04, Iersel et al. 08). For higher levels, partial answers were obtained by Iersel et al. 2008 with a polynomial time algorithm for simple networks. In this paper, we detail the first complete answer for the general case, solving a problem proposed by Jansson et al. 2004: for any $k$ fixed, it is possible to construct a minimum level-$k$ network consistent with $\mathcal{T}$, if there is any, in time $O(|\mathcal{T}|^{k+1}n^{\lfloor\frac{4k}{3}\rfloor+1})$. This is an improved result of our preliminary version presented at CPM'2009. △ Less

Submitted 23 June, 2011; v1 submitted 11 March, 2011; originally announced March 2011.

Structure and Recognition of 3,4-leaf Powers of Galled Phylogenetic Networks in Polynomial Time

Authors: Michel Habib, Thu-Hien To

Abstract: A graph is a $k$-leaf power of a tree $T$ if its vertices are leaves of $T$ and two vertices are adjacent in $T$ if and only if their distance in $T$ is at most $k$. Then $T$ is a $k$-leaf root of $G$. This notion was introduced by Nishimura, Ragde, and Thilikos [2002] motivated by the search for underlying phylogenetic trees. We study here an extension of the $k$-leaf power graph recognition prob… ▽ More A graph is a $k$-leaf power of a tree $T$ if its vertices are leaves of $T$ and two vertices are adjacent in $T$ if and only if their distance in $T$ is at most $k$. Then $T$ is a $k$-leaf root of $G$. This notion was introduced by Nishimura, Ragde, and Thilikos [2002] motivated by the search for underlying phylogenetic trees. We study here an extension of the $k$-leaf power graph recognition problem. This extension is motivated by a new biological question for the evaluation of the latteral gene transfer on a population of viruses. We allow the host graph to slightly differs from a tree and allow some cycles. In fact we study phylogenetic galled networks in which cycles are pairwise vertex disjoint. We show some structural results and propose polynomial algorithms for the cases $k=3$ and $k=4$. As a consequence, squares of galled networks can also be recognized in polynomial time. △ Less

Submitted 23 July, 2011; v1 submitted 18 December, 2010; originally announced December 2010.

Super duality and homology of unitarizable modules of Lie algebras

Authors: Po-Yi Huang, Ngau Lam, Tze-Ming To

Abstract: The u-homology formulas for unitarizable modules at negative levels over classical Lie algebras of infinite rank of types gl(n), sp(2n) and so(2n) are obtained. As a consequence, we recover the Enright's formulas for three Hermitian symmetric pairs of classical types (SU(p; q); SU(p) X SU(q)), (Sp(2n);U(n)) and (SO*(2n);U(n)). The u-homology formulas for unitarizable modules at negative levels over classical Lie algebras of infinite rank of types gl(n), sp(2n) and so(2n) are obtained. As a consequence, we recover the Enright's formulas for three Hermitian symmetric pairs of classical types (SU(p; q); SU(p) X SU(q)), (Sp(2n);U(n)) and (SO*(2n);U(n)). △ Less

Submitted 6 December, 2010; originally announced December 2010.

Comments: 16 pages

Level-k Phylogenetic Network can be Constructed from a Dense Triplet Set in Polynomial Time

Authors: Thu-Hien To, Michel Habib

Abstract: Given a dense triplet set $\mathcal{T}$, there arise two interesting questions: Does there exists any phylogenetic network consistent with $\mathcal{T}$? And if so, can we find an effective algorithm to construct one? For cases of networks of levels $k=0$ or 1 or 2, these questions were answered with effective polynomial algorithms. For higher levels $k$, partial answers were recently obtained w… ▽ More Given a dense triplet set $\mathcal{T}$, there arise two interesting questions: Does there exists any phylogenetic network consistent with $\mathcal{T}$? And if so, can we find an effective algorithm to construct one? For cases of networks of levels $k=0$ or 1 or 2, these questions were answered with effective polynomial algorithms. For higher levels $k$, partial answers were recently obtained with an $O(|\mathcal{T}|^{k+1})$ time algorithm for simple networks. In this paper we give a complete answer to the general case. The main idea is to use a special property of SN-sets in a level-k network. As a consequence, we can also find the level-k network with the minimum number of reticulations in polynomial time. △ Less

Submitted 12 January, 2009; originally announced January 2009.

Regularity of the minimizers in the composite membrane problem in R^2

Authors: Sagun Chanillo, Carlos E. Kenig, Tung TO

Abstract: We study the regularity of minimizers to the composite membrane problem in the plane (ie given a domain omega and a positive number A, smaller than the measure of omega, minimize the first Dirichlet eigenvalue for the Schrodinger operator with potential equal to a fixed multiple of the characteristic function of a subset D of omega, with measure A). We show that for minimizers, the boundary of D… ▽ More We study the regularity of minimizers to the composite membrane problem in the plane (ie given a domain omega and a positive number A, smaller than the measure of omega, minimize the first Dirichlet eigenvalue for the Schrodinger operator with potential equal to a fixed multiple of the characteristic function of a subset D of omega, with measure A). We show that for minimizers, the boundary of D is analytic. △ Less

Submitted 7 April, 2008; originally announced April 2008.

MSC Class: 35R35

A free-boundary problem for the evolution $p$-Laplacian equation with a combustion boundary condition

Authors: Tung To

Abstract: We study the existence, uniqueness and regularity of solutions of the equation $f_t = Δ_p f = \text{div} (|Df|^{p-2} Df)$ under over-determined boundary conditions $f = 0$ and $|Df| = 1$. We show that if the initial data is concave and Lipschitz with a bounded and convex support, then the problem admits a unique solution which exists until it vanishes identically. Furthermore, the free-boundary… ▽ More We study the existence, uniqueness and regularity of solutions of the equation $f_t = Δ_p f = \text{div} (|Df|^{p-2} Df)$ under over-determined boundary conditions $f = 0$ and $|Df| = 1$. We show that if the initial data is concave and Lipschitz with a bounded and convex support, then the problem admits a unique solution which exists until it vanishes identically. Furthermore, the free-boundary of the support of $f$ is smooth for all positive time. △ Less

Submitted 20 November, 2007; originally announced November 2007.

Comments: 25 pages, submitted

MSC Class: 35R35 (Primary); 35K55; 35K65 (Secondary)