arXiv subjects

Li Ding

Publications and source records attributed to Li Ding.

At least 55 records · Page 3Linked to original sources

arXiv2022

Evolving Neural Selection with Adaptive Regularization

Over-parameterization is one of the inherent characteristics of modern deep neural networks, which can often be overcome by leveraging regularization methods, such as Dropout. Usually, these methods are applied globally and all the input cases are treated equally. However, given the natural variation of the input space for real-world tasks such as image recognition and natural language understanding, it is unlikely that a fixed regularization pattern will have the same effectiveness for all the input cases. In this work, we demonstrate a method in which the selection of neurons in deep neural networks evolves, adapting to the difficulty of prediction. We propose the Adaptive Neural Selection (ANS) framework, which evolves to weigh neurons in a layer to form network variants that are suitable to handle different input cases. Experimental results show that the proposed method can significantly improve the performance of commonly-used neural network architectures on standard image recognition benchmarks. Ablation studies also validate the effectiveness and contribution of each component in the proposed framework.

cs.NE↗

arXiv2021

Delving into Deep Image Prior for Adversarial Defense: A Novel Reconstruction-based Defense Framework

Deep learning based image classification models are shown vulnerable to adversarial attacks by injecting deliberately crafted noises to clean images. To defend against adversarial attacks in a training-free and attack-agnostic manner, this work proposes a novel and effective reconstruction-based defense framework by delving into deep image prior (DIP). Fundamentally different from existing reconstruction-based defenses, the proposed method analyzes and explicitly incorporates the model decision process into our defense. Given an adversarial image, firstly we map its reconstructed images during DIP optimization to the model decision space, where cross-boundary images can be detected and on-boundary images can be further localized. Then, adversarial noise is purified by perturbing on-boundary images along the reverse direction to the adversarial image. Finally, on-manifold images are stitched to construct an image that can be correctly predicted by the victim classifier. Extensive experiments demonstrate that the proposed method outperforms existing state-of-the-art reconstruction-based methods both in defending white-box attacks and defense-aware attacks. Moreover, the proposed method can maintain a high visual quality during adversarial image reconstruction.

cs.CV↗

arXiv2021

Deep Weakly Supervised Positioning

PoseNet can map a photo to the position where it is taken, which is appealing in robotics. However, training PoseNet requires full supervision, where ground truth positions are non-trivial to obtain. Can we train PoseNet without knowing the ground truth positions for each observation? We show that this is possible via constraint-based weak-supervision, leading to the proposed framework: DeepGPS. Particularly, using wheel-encoder-estimated distances traveled by a robot along random straight line segments as constraints between PoseNet outputs, DeepGPS can achieve a relative positioning error of less than 2%. Moreover, training DeepGPS can be done as auto-calibration with almost no human attendance, which is more attractive than its competing methods that typically require careful and expert-level manual calibration. We conduct various experiments on simulated and real datasets to demonstrate the general applicability, effectiveness, and accuracy of DeepGPS, and perform a comprehensive analysis of its robustness. Our code is available at https://ai4ce.github.io/DeepGPS/.

cs.RO↗

arXiv2020

Perception Matters: Exploring Imperceptible and Transferable Anti-forensics for GAN-generated Fake Face Imagery Detection

Recently, generative adversarial networks (GANs) can generate photo-realistic fake facial images which are perceptually indistinguishable from real face photos, promoting research on fake face detection. Though fake face forensics can achieve high detection accuracy, their anti-forensic counterparts are less investigated. Here we explore more \textit{imperceptible} and \textit{transferable} anti-forensics for fake face imagery detection based on adversarial attacks. Since facial and background regions are often smooth, even small perturbation could cause noticeable perceptual impairment in fake face images. Therefore it makes existing adversarial attacks ineffective as an anti-forensic method. Our perturbation analysis reveals the intuitive reason of the perceptual degradation issue when directly applying existing attacks. We then propose a novel adversarial attack method, better suitable for image anti-forensics, in the transformed color domain by considering visual perception. Simple yet effective, the proposed method can fool both deep learning and non-deep learning based forensic detectors, achieving higher attack success rate and significantly improved visual quality. Specially, when adversaries consider imperceptibility as a constraint, the proposed anti-forensic method can improve the average attack success rate by around 30\% on fake face images over two baseline attacks. \textit{More imperceptible} and \textit{more transferable}, the proposed method raises new security concerns to fake face imagery detection. We have released our code for public use, and hopefully the proposed method can be further explored in related forensic applications as an anti-forensic benchmark.

cs.CV↗

arXiv2020

A Novel Deep Learning Pipeline for Retinal Vessel Detection in Fluorescein Angiography

While recent advances in deep learning have significantly advanced the state of the art for vessel detection in color fundus (CF) images, the success for detecting vessels in fluorescein angiography (FA) has been stymied due to the lack of labeled ground truth datasets. We propose a novel pipeline to detect retinal vessels in FA images using deep neural networks that reduces the effort required for generating labeled ground truth data by combining two key components: cross-modality transfer and human-in-the-loop learning. The cross-modality transfer exploits concurrently captured CF and fundus FA images. Binary vessels maps are first detected from CF images with a pre-trained neural network and then are geometrically registered with and transferred to FA images via robust parametric chamfer alignment to a preliminary FA vessel detection obtained with an unsupervised technique. Using the transferred vessels as initial ground truth labels for deep learning, the human-in-the-loop approach progressively improves the quality of the ground truth labeling by iterating between deep-learning and labeling. The approach significantly reduces manual labeling effort while increasing engagement. We highlight several important considerations for the proposed methodology and validate the performance on three datasets. Experimental results demonstrate that the proposed pipeline significantly reduces the annotation effort and the resulting deep learning methods outperform prior existing FA vessel detection methods by a significant margin. A new public dataset, RECOVERY-FA19, is introduced that includes high-resolution ultra-widefield images and accurately labeled ground truth binary vessel maps.

cs.CV↗

arXiv2019

MIT Advanced Vehicle Technology Study: Large-Scale Naturalistic Driving Study of Driver Behavior and Interaction with Automation

For the foreseeble future, human beings will likely remain an integral part of the driving task, monitoring the AI system as it performs anywhere from just over 0% to just under 100% of the driving. The governing objectives of the MIT Autonomous Vehicle Technology (MIT-AVT) study are to (1) undertake large-scale real-world driving data collection that includes high-definition video to fuel the development of deep learning based internal and external perception systems, (2) gain a holistic understanding of how human beings interact with vehicle automation technology by integrating video data with vehicle state data, driver characteristics, mental models, and self-reported experiences with technology, and (3) identify how technology and other factors related to automation adoption and use can be improved in ways that save lives. In pursuing these objectives, we have instrumented 23 Tesla Model S and Model X vehicles, 2 Volvo S90 vehicles, 2 Range Rover Evoque, and 2 Cadillac CT6 vehicles for both long-term (over a year per driver) and medium term (one month per driver) naturalistic driving data collection. Furthermore, we are continually developing new methods for analysis of the massive-scale dataset collected from the instrumented vehicle fleet. The recorded data streams include IMU, GPS, CAN messages, and high-definition video streams of the driver face, the driver cabin, the forward roadway, and the instrument cluster (on select vehicles). The study is on-going and growing. To date, we have 122 participants, 15,610 days of participation, 511,638 miles, and 7.1 billion video frames. This paper presents the design of the study, the data collection hardware, the processing of the data, and the computer vision algorithms currently being used to extract actionable knowledge from the data.

cs.CY↗

arXiv2019

Object as Distribution

Object detection is a critical part of visual scene understanding. The representation of the object in the detection task has important implications on the efficiency and feasibility of annotation, robustness to occlusion, pose, lighting, and other visual sources of semantic uncertainty, and effectiveness in real-world applications (e.g., autonomous driving). Popular object representations include 2D and 3D bounding boxes, polygons, splines, pixels, and voxels. Each have their strengths and weakness. In this work, we propose a new representation of objects based on the bivariate normal distribution. This distribution-based representation has the benefit of robust detection of highly-overlapping objects and the potential for improved downstream tracking and instance segmentation tasks due to the statistical representation of object edges. We provide qualitative evaluation of this representation for the object detection task and quantitative evaluation of its use in a baseline algorithm for the instance segmentation task.

cs.CV↗

arXiv2019

DeepMapping: Unsupervised Map Estimation From Multiple Point Clouds

We propose DeepMapping, a novel registration framework using deep neural networks (DNNs) as auxiliary functions to align multiple point clouds from scratch to a globally consistent frame. We use DNNs to model the highly non-convex mapping process that traditionally involves hand-crafted data association, sensor pose initialization, and global refinement. Our key novelty is that "training" these DNNs with properly defined unsupervised losses is equivalent to solving the underlying registration problem, but less sensitive to good initialization than ICP. Our framework contains two DNNs: a localization network that estimates the poses for input point clouds, and a map network that models the scene structure by estimating the occupancy status of global coordinates. This allows us to convert the registration problem to a binary occupancy classification, which can be solved efficiently using gradient-based optimization. We further show that DeepMapping can be readily extended to address the problem of Lidar SLAM by imposing geometric constraints between consecutive point clouds. Experiments are conducted on both simulated and real datasets. Qualitative and quantitative comparisons demonstrate that DeepMapping often enables more robust and accurate global registration of multiple point clouds than existing techniques. Our code is available at https://ai4ce.github.io/DeepMapping/.

cs.CV↗

arXiv2019

Value of Temporal Dynamics Information in Driving Scene Segmentation

Semantic scene segmentation has primarily been addressed by forming representations of single images both with supervised and unsupervised methods. The problem of semantic segmentation in dynamic scenes has begun to recently receive attention with video object segmentation approaches. What is not known is how much extra information the temporal dynamics of the visual scene carries that is complimentary to the information available in the individual frames of the video. There is evidence that the human visual system can effectively perceive the scene from temporal dynamics information of the scene's changing visual characteristics without relying on the visual characteristics of individual snapshots themselves. Our work takes steps to explore whether machine perception can exhibit similar properties by combining appearance-based representations and temporal dynamics representations in a joint-learning problem that reveals the contribution of each toward successful dynamic scene segmentation. Additionally, we provide the MIT Driving Scene Segmentation dataset, which is a large-scale full driving scene segmentation dataset, densely annotated for every pixel and every one of 5,000 video frames. This dataset is intended to help further the exploration of the value of temporal dynamics information for semantic segmentation in video.

cs.CV↗

arXiv2018

Arguing Machines: Human Supervision of Black Box AI Systems That Make Life-Critical Decisions

We consider the paradigm of a black box AI system that makes life-critical decisions. We propose an "arguing machines" framework that pairs the primary AI system with a secondary one that is independently trained to perform the same task. We show that disagreement between the two systems, without any knowledge of underlying system design or operation, is sufficient to arbitrarily improve the accuracy of the overall decision pipeline given human supervision over disagreements. We demonstrate this system in two applications: (1) an illustrative example of image classification and (2) on large-scale real-world semi-autonomous driving data. For the first application, we apply this framework to image classification achieving a reduction from 8.0% to 2.8% top-5 error on ImageNet. For the second application, we apply this framework to Tesla Autopilot and demonstrate the ability to predict 90.4% of system disengagements that were labeled by human annotators as challenging and needing human supervision.

cs.AI↗

arXiv2018

Self Healing Interconnects with a Near Plastically Stretchable Heal

Flexible electronic systems such as roll up displays, wearable devices etc. promise exciting possibilities that could change the way humans interact with the environment. However, they suffer from poor reliability of interconnects and devices. Interconnects on flex are prone to open circuit failures due to mechanical stress, electrostatic discharge and environmental degradation. Passive approaches such as the use of stretchable conductors and novel geometries improve their response to mechanical stress but cannot salvage the interconnect if a fault were to occur. Active approaches using self healing techniques can repair a fault and have been demonstrated using methods that either use relatively rare materials, change conventional interconnect fabrication processes, only address faults due to mechanical stress or do not permit stretching. In this work we discuss a self healing technique that overcomes these limitations and demonstrate heals having metallic conductivity and nearly plastic stretchability. This is achieved using a dispersion of conductive particles in an insulating fluid encapsulated over the interconnect. Healing is automatically triggered by the electric field appearing in the open gap of a failed interconnect, irrespective of the cause of failure. The field polarizes the conductive particles causing them aggregate and chain up to bridge the gap and repair the fault. Using copper-silicone oil dispersions, we show self healing interconnects with the stretchable heal having conductivity of about $5 \times 10^{5}$ Sm and allowing strains from 12 to 60. Previously, stretchable interconnects used materials other than copper. Here we effectively show self healing, stretchable copper. This work promises high speed, self healing and stretchable interconnects on flex thereby improving system reliability.

cond-mat.soft↗

arXiv2018

Weakly-Supervised Action Segmentation with Iterative Soft Boundary Assignment

In this work, we address the task of weakly-supervised human action segmentation in long, untrimmed videos. Recent methods have relied on expensive learning models, such as Recurrent Neural Networks (RNN) and Hidden Markov Models (HMM). However, these methods suffer from expensive computational cost, thus are unable to be deployed in large scale. To overcome the limitations, the keys to our design are efficiency and scalability. We propose a novel action modeling framework, which consists of a new temporal convolutional network, named Temporal Convolutional Feature Pyramid Network (TCFPN), for predicting frame-wise action labels, and a novel training strategy for weakly-supervised sequence modeling, named Iterative Soft Boundary Assignment (ISBA), to align action sequences and update the network in an iterative fashion. The proposed framework is evaluated on two benchmark datasets, Breakfast and Hollywood Extended, with four different evaluation metrics. Extensive experimental results show that our methods achieve competitive or superior performance to state-of-the-art methods.

cs.CV↗

arXiv2017

TricorNet: A Hybrid Temporal Convolutional and Recurrent Network for Video Action Segmentation

Action segmentation as a milestone towards building automatic systems to understand untrimmed videos has received considerable attention in the recent years. It is typically being modeled as a sequence labeling problem but contains intrinsic and sufficient differences than text parsing or speech processing. In this paper, we introduce a novel hybrid temporal convolutional and recurrent network (TricorNet), which has an encoder-decoder architecture: the encoder consists of a hierarchy of temporal convolutional kernels that capture the local motion changes of different actions; the decoder is a hierarchy of recurrent neural networks that are able to learn and memorize long-term action dependencies after the encoding stage. Our model is simple but extremely effective in terms of video sequence labeling. The experimental results on three public action segmentation datasets have shown that the proposed model achieves superior performance over the state of the art.

cs.CV↗

arXiv2008

Production of the neutral top-pion $π_{t}^{0}$ in association with a high-$p_{T}$ jet at the $LHC$

In the framework of the topcolor-assisted technicolor $(TC2)$ model, we study production of the neutral top-pion $π_{t}^{0}$ in association with a high-$p_{T}$ jet at the $LHC$, which proceeds via the partonic processes $gg\longrightarrow π_{t}^{0}g$, $gq\longrightarrow π_{t}^{0}q$, $q\bar{q}\longrightarrow π_{t}^{0}g$, $gb(\bar{b})\longrightarrow π_{t}^{0}b(\bar{b})$, and $b\bar{b}\longrightarrow π_{t}^{0}g$. We find that it is very challenging to detect the neutral top-pion $π_{t}^{0}$ via the process $pp\longrightarrow π_{t}^{0}+jet+X\to t\bar{t}+jet+X$, while the possible signatures of $π_{t}^{0}$ might be detected via the process $pp\longrightarrow π_{t}^{0}+jet+X\to(\bar{t}c+t\bar{c})+jet+X$ at the $LHC$.

hep-ph↗

arXiv2008

Photon-induced production of the mirror quarks from the $LHT$ model at the $LHC$

The photon-induced processes at the $LHC$ provide clean experimental conditions due to absence of the proton remnants, which might produce complementary and interesting results for tests of the standard model and for searching of new physics. In the context of the littlest $Higgs$ model with T-parity, we consider the photon-induced production of the mirror quarks at the $LHC$. The cross sections for various production channels are calculated and a simply phenomenology analysis is performed by assuming leptonic decays.

hep-ph↗

arXiv2008

Production of mirror fermions via $eγ$ and $ep$ collisions in the littlest Higgs model with T-parity

One of the important features of the littlest Higgs model with T-parity, called the $LHT$ model, is that it introduces the mirror fermions, which are the T-parity partners of the standard model fermions. In this paper, we discuss production of the mirror quark associated with mirror neutrino via $eγ$ and $ep$ collisions. We find that, in wide range of the parameter space, the mirror quark can be copiously produced at the International $e^{+}e^{-}$ Linear Collider $(ILC)$ and $ep$ collider $(THERA)$ experiments. The production rates of certain signal events, which are related the main two-body decay modes of the mirror quark, are also calculated.

hep-ph↗

arXiv2008

The new charged gauge boson $W'$ and the subprocess $eq\toνq'$ at $e^{+}e^{-}$ and $ep$ colliders

In the framework of the little Higgs models and the three-site Higgsless model, we discuss the contributions of the new charged gauge boson $W'$ to the process $eq\toνq'$ and the possibility of detecting $W'$ via this process in future high energy linear $e^{+}e^{-}$ collider $(ILC)$ and $ep$ collider $(THERA)$ experiments. Our numerical results show that the process $eq\toνq'$ is rather sensitive to the coupling $W'ff'$ and one can use this process to distinguish different new physics models in future $ILC$ and $THERA$ experiments.

hep-ph↗

arXiv2008

Top quark chromomagnetic dipole moment in the littlest Higgs model with T-parity

The littlest Higgs model with T-parity, which is called $LHT$ model, predicts the existence of the new particles, such as heavy top quark, heavy gauge bosons, and mirror fermions. We calculate the one-loop contributions of these new particles to the top quark chromomagnetic dipole moment $(CMDM)$ $ΔK$. We find that the contribution of the $LHT$ model is one order of magnitude smaller than the standard model prediction value.

hep-ph↗