We show that, for each of five datasets of increasing complexity, certain training samples are more informative of class membership than others. These samples can be identified a priori to training by analyzing their position in reduced dimensional space relative to the classes' centroids. Specifically, we demonstrate that samples nearer the classes' centroids are less informative than those that are furthest from it. For all five datasets, we show that there is no statistically significant difference between training on the entire training set and when excluding up to 2% of the data nearest to each class's centroid.
We present a Conformer-based end-to-end neural diarization (EEND) model that uses both acoustic input and features derived from an automatic speech recognition (ASR) model. Two categories of features are explored: features derived directly from ASR output (phones, position-in-word and word boundaries) and features derived from a lexical speaker change detection model, trained by fine-tuning a pretrained BERT model on the ASR output. Three modifications to the Conformer-based EEND architecture are proposed to incorporate the features. First, ASR features are concatenated with acoustic features. Second, we propose a new attention mechanism called contextualized self-attention that utilizes ASR features to build robust speaker representations. Finally, multi-task learning is used to train the model to minimize classification loss for the ASR features along with diarization loss. Experiments on the two-speaker English conversations of Switchboard+SRE data sets show that multi-task learning with position-in-word information is the most effective way of utilizing ASR features, reducing the diarization error rate (DER) by 20% relative to the baseline.
Vision based solutions for the localization of vehicles have become popular recently. We employ an image retrieval based visual localization approach. The database images are kept with GPS coordinates and the location of the retrieved database image serves as an approximate position of the query image. We show that localization can be performed via descriptors solely extracted from semantically segmented images. It is reliable especially when the environment is subjected to severe illumination and seasonal changes. Our experiments reveal that the localization performance of a semantic descriptor can increase up to the level of state-of-the-art RGB image based methods.
We study the problem of score and rank monotonicity for spectral ranking methods, such as eigenvector centrality and PageRank, in the case of undirected networks. Score monotonicity means that adding an edge increases the score at both ends of the edge. Rank monotonicity means that adding an edge improves the relative position of both ends of the edge with respect to the remaining nodes. It is known that common spectral rankings are both score and rank monotone on directed, strongly connected graphs. We show that, surprisingly, the situation is very different for undirected graphs, and in particular that PageRank is neither score nor rank monotone.
Meaning is defined by the company it keeps. However, company is two-fold: It's based on the identity of tokens and also on their position (topology). We argue that a position-centric perspective is more general and useful. The classic MLM and CLM objectives in NLP are easily phrased as position predictions over the whole vocabulary. Adapting the relative position encoding paradigm in NLP to create relative labels for self-supervised learning, we seek to show superior pre-training judged by performance on downstream tasks.
Mechatronic systems have increasingly high performance requirements for motion control. The low-frequency contribution of the flexible dynamics, i.e. the compliance, should be compensated for by means of snap feedforward to achieve high accuracy. Position-dependent compliance, which often occurs in motion systems, requires the snap feedforward parameter to be modeled as a function of position. Position-dependent compliance is compensated for by using a Gaussian process to model the snap feedforward parameter as a continuous function of position. A simulation of a flexible beam shows that a significant performance increase is achieved when using the Gaussian process snap feedforward parameter to compensate for position-dependent compliance.
This paper addresses high-resolution vehicle positioning and tracking. In recent work, it was shown that a fleet of independent but neighboring vehicles can cooperate for the task of localization by capitalizing on the existence of common surrounding reflectors, using the concept of Team Channel-SLAM. This approach exploits an initial (e.g. GPS-based) vehicle position information and allows subsequent tracking of vehicles by exploiting the shared nature of virtual transmitters associated to the reflecting surfaces. In this paper, we show that the localization can be greatly enhanced by joint sensing and mapping of reflecting surfaces. To this end, we propose a combined approach coined Team Channel-SLAM Evolution (TCSE) which exploits the intertwined relation between (i) the position of virtual transmitters, (ii) the shape of reflecting surfaces, and (iii) the paths described by the radio propagation rays, in order to achieve high-resolution vehicle localization. Overall, TCSE yields a complete picture of the trajectories followed by dominant paths together with a mapping of reflecting surfaces. While joint localization and mapping is a well researched topic within robotics using inputs such as radar and vision, this paper is first to demonstrate such an approach within mobile networking framework based on radio data.
In this paper, a machine learning based approach is introduced to estimate pendubot angular position from its captured images. Initially, a baseline algorithm is introduced to estimate the angle using conventional image processing techniques. The baseline algorithm performs well for the cases that the pendubot is not moving fast. However, when moving quickly due to a free fall, the pendubot appears as a blurred object in the captured image in a way that the baseline algorithm fails to estimate the angle. Consequently, a Deep Neural Network (DNN) based algorithm is introduced to cope with this challenge. The approach relies on the concept of transfer learning to allow the training of the DNN on a very small fine-tuning dataset. The base algorithm is used to create the ground truth labels of the fine-tuning dataset. Experimental results on the held-out evaluation set show that the proposed approach achieves a median absolute error of 0.02 and 0.06 degrees for the sharp and blurry images respectively.
Computational feasibility is a widespread concern that guides the framing and modeling of biological and artificial intelligence. The specification of cognitive system capacities is often shaped by unexamined intuitive assumptions about the search space and complexity of a subcomputation. However, a mistaken intuition might make such initial conceptualizations misleading for what empirical questions appear relevant later on. We undertake here computational-level modeling and complexity analyses of segmentation - a widely hypothesized subcomputation that plays a requisite role in explanations of capacities across domains - as a case study to show how crucial it is to formally assess these assumptions. We mathematically prove two sets of results regarding hardness and search space size that may run counter to intuition, and position their implications with respect to existing views on the subcapacity.
We propose a novel positional encoding for learning graph on Transformer architecture. Existing approaches either linearize a graph to encode absolute position in the sequence of nodes, or encode relative position with another node using bias terms. The former loses preciseness of relative position from linearization, while the latter loses a tight integration of node-edge and node-topology interaction. To overcome the weakness of the previous approaches, our method encodes a graph without linearization and considers both node-topology and node-edge interaction. We name our method Graph Relative Positional Encoding dedicated to graph representation learning. Experiments conducted on various graph datasets show that the proposed method outperforms previous approaches significantly. Our code is publicly available at https://github.com/lenscloth/GRPE.
This work investigates a multibeam system for joint sensing and communication (JSC) based on multiple-input multiple-output (MIMO) 5G new radio (NR) waveforms. In particular, we consider a base station (BS) acting as a monostatic sensor that estimates the range, speed, and direction of arrival (DoA) of multiple targets via beam scanning using a fraction of the transmitted power. The target position is then obtained via range and DoA estimation. We derive the sensing performance in terms of probability of detection and root mean squared error (RMSE) of position and velocity estimation of a target under line-of-sight (LOS) conditions. Furthermore, we evaluate the system performance when multiple targets are present, using the optimal sub-pattern assignment (OSPA) metric. Finally, we provide an in-depth investigation of the dominant factors that affect performance, including the fraction of power reserved for sensing.
We implement a data assimilation framework for integrating ice surface and terminus position observations into a numerical ice-flow model. The model uses the well-known shallow shelf approximation (SSA) coupled to a level set method to capture ice motion and changes in the glacier geometry. The level set method explicitly tracks the evolving ice-atmosphere and ice-ocean boundaries for a marine outlet glacier. We use an Ensemble Transform Kalman Filter to assimilate observations of ice surface elevation and lateral ice extent by updating the level set function that describes the ice interface. Numerical experiments on an idealized marine-terminating glacier demonstrate the effectiveness of our data assimilation approach for tracking seasonal and multi-year glacier advance and retreat cycles. The model is also applied to simulate Helheim Glacier, a major tidewater-terminating glacier of the Greenland Ice Sheet that has experienced a recent history of rapid retreat. By assimilating observations from remotely-sensed surface elevation profiles we are able to more accurately track the migrating glacier terminus and glacier surface changes. These results support the use of data assimilation methodologies for obtaining more accurate predictions of short-term ice sheet dynamics.
Kalman filter-based Inertial Navigation System (INS) is a reliable and efficient method to estimate the position of a pedestrian indoors. Classical INS-based methodology which is called IEZ (INS-EKF-ZUPT) makes use of an Extended Kalman Filter (EKF), a Zero velocity UPdaTing (ZUPT) to calculate the position and attitude of a person. However, heading error which is a key factor of the whole Pedestrian Dead Reckoning (PDR) system is unobservable for IEZ-based PDR system. To minimize the error, Electronic Com-pass (EC) algorithm becomes a valid method. But magnetic disturbance may have a big negative effect on it. In this paper, the Quasi-static Magnetic field Detection (QMD) method is proposed to detect the pure magnetic field and then selects EC algorithm or Heuristic heading Drift Reduction algorithm (HDR) according to the detection result, which implements the complementation of the two methods. Meanwhile, the QMD, EC, and HDR algorithms are integrated into the IEZ framework to form a new PDR solution which is named Advanced IEZ (AIEZ).
Mechatronic systems have increasingly stringent performance requirements for motion control, leading to a situation where many factors, such as position-dependency, cannot be neglected in feedforward control. The aim of this paper is to compensate for position-dependent effects by modeling feedforward parameters as a function of position. A framework to model and identify feedforward parameters as a continuous function of position is developed by combining Gaussian processes and feedforward parameter learning techniques. The framework results in a fully data-driven approach, which can be readily implemented for industrial control applications. The framework is experimentally validated and shows a significant performance increase on a commercial wire bonder.
Progress in 3D object understanding has relied on manually canonicalized shape datasets that contain instances with consistent position and orientation (3D pose). This has made it hard to generalize these methods to in-the-wild shapes, eg., from internet model collections or depth sensors. ConDor is a self-supervised method that learns to Canonicalize the 3D orientation and position for full and partial 3D point clouds. We build on top of Tensor Field Networks (TFNs), a class of permutation- and rotation-equivariant, and translation-invariant 3D networks. During inference, our method takes an unseen full or partial 3D point cloud at an arbitrary pose and outputs an equivariant canonical pose. During training, this network uses self-supervision losses to learn the canonical pose from an un-canonicalized collection of full and partial 3D point clouds. ConDor can also learn to consistently co-segment object parts without any supervision. Extensive quantitative results on four new metrics show that our approach outperforms existing methods while enabling new applications such as operation on depth images and annotation transfer.
In this paper we present GroupGaze. It is a tool that can be used to calculate the gaze direction and the gaze position of whole groups. GroupGazer calculates the gaze direction of every single person in the image and allows to map these gaze vectors to a projection like a projector. In addition to the person-specific gaze direction, the person affiliation of each gaze vector is stored based on the position in the image. Also, it is possible to save the group attention after a calibration. The software is free to use and requires a simple webcam as well as an NVIDIA GPU and the operating system Windows or Linux. Link: https://es-cloud.cs.uni-tuebingen.de/d/8e2ab8c3fdd444e1a135/?p=%2FGroupGazer&mode=list
Incorporating geometric transformations that reflect the relative position changes between an observer and an object into computer vision and deep learning models has attracted much attention in recent years. However, the existing proposals mainly focus on the affine transformation that is insufficient to reflect such geometric position changes. Furthermore, current solutions often apply a neural network module to learn a single transformation matrix, which not only ignores the importance of multi-view analysis but also includes extra training parameters from the module apart from the transformation matrix parameters that increase the model complexity. In this paper, a perspective transformation layer is proposed in the context of deep learning. The proposed layer can learn homography, therefore reflecting the geometric positions between observers and objects. In addition, by directly training its transformation matrices, a single proposed layer can learn an adjustable number of multiple viewpoints without considering module parameters. The experiments and evaluations confirm the superiority of the proposed layer.
Facial micro-expressions (MEs) are involuntary facial motions revealing peoples real feelings and play an important role in the early intervention of mental illness, the national security, and many human-computer interaction systems. However, existing micro-expression datasets are limited and usually pose some challenges for training good classifiers. To model the subtle facial muscle motions, we propose a robust micro-expression recognition (MER) framework, namely muscle motion-guided network (MMNet). Specifically, a continuous attention (CA) block is introduced to focus on modeling local subtle muscle motion patterns with little identity information, which is different from most previous methods that directly extract features from complete video frames with much identity information. Besides, we design a position calibration (PC) module based on the vision transformer. By adding the position embeddings of the face generated by PC module at the end of the two branches, the PC module can help to add position information to facial muscle motion pattern features for the MER. Extensive experiments on three public micro-expression datasets demonstrate that our approach outperforms state-of-the-art methods by a large margin.
The use of the LoRa communication protocol in a new generation of transceivers is attractive for search and rescue(SaR) procedures because they can operate in harsh environmentscovering vast areas while maintaining a low power consumption.The possibility of wearing helmets equipped with LoRa-radiosand installing LoRa transceivers in unmanned aerial vehicles (UAVs) will accelerate the localization of the targets, probably unconscious. In this paper, the achievable communication ranges of such links are theoretically and experimentally evaluated by considering the possible positions of the helmet wearer (standing or lying) on a flat field, representing a simple SaR scenario.Simulations and experimental tests demonstrated that, for the standing position, the ground-bounce multi-path produces strong fluctuations of the received power versus the Tx-Rx distances. Such fluctuations can be kept confined within 100 m from the target by lowering the UAV altitude. Instead, for a more critical lying position, the received power profile is monotonic and nearly insensitive to the posture. For all the considered cases, the signal emitted by the body-worn transceiver can be exploited to localize the helmet wearer based on its strength, and it is theoretically detectable by the UAV radio up to 5 km on flat terrain.
We present a novel graph-based approach for labeling the anatomical branches of a given airway tree segmentation. The proposed method formulates airway labeling as a branch classification problem in the airway tree graph, where branch features are extracted using convolutional neural networks (CNN) and enriched using graph neural networks. Our graph neural network is structure-aware by having each node aggregate information from its local neighbors and position-aware by encoding node positions in the graph. We evaluated the proposed method on 220 airway trees from subjects with various severity stages of Chronic Obstructive Pulmonary Disease (COPD). The results demonstrate that our approach is computationally efficient and significantly improves branch classification performance than the baseline method. The overall average accuracy of our method reaches 91.18\% for labeling all 18 segmental airway branches, compared to 83.83\% obtained by the standard CNN method. We published our source code at https://github.com/DIAGNijmegen/spgnn. The proposed algorithm is also publicly available at https://grand-challenge.org/algorithms/airway-anatomical-labeling/.