Multi-modal Ads Video Understanding Challenge is the first grand challenge aiming to comprehensively understand ads videos. Our challenge includes two tasks: video structuring in the temporal dimension and multi-modal video classification. It asks th
e participants to accurately predict both the scene boundaries and the multi-label categories of each scene based on a fine-grained and ads-related category hierarchy. Therefore, our task has four distinguishing features from previous ones: ads domain, multi-modal information, temporal segmentation, and multi-label classification. It will advance the foundation of ads video understanding and have a significant impact on many ads applications like video recommendation. This paper presents an overview of our challenge, including the background of ads videos, an elaborate description of task and dataset, evaluation protocol, and our proposed baseline. By ablating the key components of our baseline, we would like to reveal the main challenges of this task and provide useful guidance for future research of this area. In this paper, we give an extended version of our challenge overview. The dataset will be publicly available at https://algo.qq.com/.
In particle physics, semi-supervised machine learning is an attractive option to reduce model dependencies searches beyond the Standard Model. When utilizing semi-supervised techniques in training machine learning models in the search for bosons at t
he Large Hadron Collider, the over-training of the model must be investigated. Internal fluctuations of the phase space and bias in training can cause semi-supervised models to label false signals within the phase space due to over-fitting. The issue of false signal generation in semi-supervised models has not been fully analyzed and therefore utilizing a toy Monte Carlo model, the probability of such situations occurring must be quantified. This investigation of $Zgamma$ resonances is performed using a pure background Monte Carlo sample. Through unique pure background samples extracted to mimic ATLAS data in a background-plus-signal region, multiple runs enable the probability of these fake signals occurring due to over-training to be thoroughly investigated.
Non-collinear antiferromagnets exhibits richer magneto-transport properties due to the topologically nontrivial spin structure they possess compared to conventional nonmagnetic materials, which allows us to manipulate the charge-spin conversion more
freely by taking advantage of the chirality. In this work, we explore the unconventional spin orbit torque of L1$_2$-ordered Mn$_3$Pt with a triangular spin structure. We observed an unconventional spin orbit torque along the $mathbf{x}$-direction for the (001)-oriented L1$_2$ Mn$_3$Pt, and found that it has a unique sign reversal behavior relative to the crystalline orientation. This generation of unconventional spin orbit torque for L1$_2$-ordered Mn$_3$Pt can be interpreted as stemming from the magnetic spin Hall effect. This report help clarify the correlation between the topologically nontrivial spin structure and charge-spin conversion in non-collinear antiferromagnets.
We present an algorithm to compute planar linkage topology and geometry, given a user-specified end-effector trajectory. Planar linkage structures convert rotational or prismatic motions of a single actuator into an arbitrarily complex periodic motio
n, refined{which is an important component when building low-cost, modular robots, mechanical toys, and foldable structures in our daily lives (chairs, bikes, and shelves). The design of such structures require trial and error even for experienced engineers. Our research provides semi-automatic methods for exploring novel designs given high-level specifications and constraints.} We formulate this problem as a non-smooth numerical optimization with quadratic objective functions and non-convex quadratic constraints involving mixed-integer decision variables (MIQCQP). We propose and compare three approximate algorithms to solve this problem: mixed-integer conic-programming (MICP), mixed-integer nonlinear programming (MINLP), and simulated annealing (SA). We evaluated these algorithms searching for planar linkages involving $10-14$ rigid links. Our results show that the best performance can be achieved by combining MICP and MINLP, leading to a hybrid algorithm capable of finding the planar linkages within a couple of hours on a desktop machine, which significantly outperforms the SA baseline in terms of optimality. We highlight the effectiveness of our optimized planar linkages by using them as legs of a walking robot.
Graph convolutional networks (GCNs) achieve promising performance for skeleton-based action recognition. However, in most GCN-based methods, the spatial-temporal graph convolution is strictly restricted by the graph topology while only captures the s
hort-term temporal context, thus lacking the flexibility of feature extraction. In this work, we present a novel architecture, named Graph Convolutional skeleton Transformer (GCsT), which addresses limitations in GCNs by introducing Transformer. Our GCsT employs all the benefits of Transformer (i.e. dynamical attention and global context) while keeps the advantages of GCNs (i.e. hierarchy and local topology structure). In GCsT, the spatial-temporal GCN forces the capture of local dependencies while Transformer dynamically extracts global spatial-temporal relationships. Furthermore, the proposed GCsT shows stronger expressive capability by adding additional information present in skeleton sequences. Incorporating the Transformer allows that information to be introduced into the model almost effortlessly. We validate the proposed GCsT by conducting extensive experiments, which achieves the state-of-the-art performance on NTU RGB+D, NTU RGB+D 120 and Northwestern-UCLA datasets.
Designing an optimal deep neural network for a given task is important and challenging in many machine learning applications. To address this issue, we introduce a self-adaptive algorithm: the adaptive network enhancement (ANE) method, written as loo
ps of the form train, estimate and enhance. Starting with a small two-layer neural network (NN), the step train is to solve the optimization problem at the current NN; the step estimate is to compute a posteriori estimator/indicators using the solution at the current NN; the step enhance is to add new neurons to the current NN. Novel network enhancement strategies based on the computed estimator/indicators are developed in this paper to determine how many new neurons and when a new layer should be added to the current NN. The ANE method provides a natural process for obtaining a good initialization in training the current NN; in addition, we introduce an advanced procedure on how to initialize newly added neurons for a better approximation. We demonstrate that the ANE method can automatically design a nearly minimal NN for learning functions exhibiting sharp transitional layers as well as discontinuous solutions of hyperbolic partial differential equations.
This report describes our submission to the track 1 and track 2 of the VoxCeleb Speaker Recognition Challenge 2021 (VoxSRC 2021). Both track 1 and track 2 share the same speaker verification system, which only uses VoxCeleb2-dev as our training set.
This report explores several parts, including data augmentation, network structures, domain-based large margin fine-tuning, and back-end refinement. Our system is a fusion of 9 models and achieves first place in these two tracks of VoxSRC 2021. The minDCF of our submission is 0.1034, and the corresponding EER is 1.8460%.
In this paper, we consider the design of a multiple-input multiple-output (MIMO) transmitter which simultaneously functions as a MIMO radar and a base station for downlink multiuser communications. In addition to a power constraint, we require the co
variance of the transmit waveform be equal to a given optimal covariance for MIMO radar, to guarantee the radar performance. With this constraint, we formulate and solve the signal-to-interference-plus-noise ratio (SINR) balancing problem for multiuser transmit beamforming via convex optimization. Considering that the interference cannot be completely eliminated with this constraint, we introduce dirty paper coding (DPC) to further cancel the interference, and formulate the SINR balancing and sum rate maximization problem in the DPC regime. Although both of the two problems are non-convex, we show that they can be reformulated to convex optimizations via the Lagrange and downlink-uplink duality. In addition, we propose gradient projection based algorithms to solve the equivalent dual problem of SINR balancing, in both transmit beamforming and DPC regimes. The simulation results demonstrate significant performance improvement of DPC over transmit beamforming, and also indicate that the degrees of freedom for the communication transmitter is restricted by the rank of the covariance.
One-stage object detectors rely on a point feature to predict the detection results. However, the point feature often lacks the information of the whole object, thereby leading to a misalignment between the object and the point feature. Meanwhile, th
e classification and regression tasks are sensitive to different object regions, but their features are spatially aligned. Both of these two problems hinder the detection performance. In order to solve these two problems, we propose a simple and plug-in operator that can generate aligned and disentangled features for each task, respectively, without breaking the fully convolutional manner. By predicting two task-aware point sets that are located in each sensitive region, the proposed operator can align the point feature with the object and disentangle the two tasks from the spatial dimension. We also reveal an interesting finding of the opposite effect of the long-range skip connection for classification and regression. On the basis of the Object-Aligned and Task-disentangled operator (OAT), we propose OAT-Net, which explicitly exploits point-set features for accurate detection results. Extensive experiments on the MS-COCO dataset show that OAT can consistently boost different state-of-the-art one-stage detectors by $sim$2 AP. Notably, OAT-Net with Res2Net-101-DCN backbone achieves 53.7 AP on the COCO test-dev.
The success of deep neural networks (DNNs) haspromoted the widespread applications of person re-identification (ReID). However, ReID systems inherit thevulnerability of DNNs to malicious attacks of visually in-conspicuous adversarial perturbations. D
etection of adver-sarial attacks is, therefore, a fundamental requirement forrobust ReID systems. In this work, we propose a Multi-Expert Adversarial Attack Detection (MEAAD) approach toachieve this goal by checking context inconsistency, whichis suitable for any DNN-based ReID systems. Specifically,three kinds of context inconsistencies caused by adversar-ial attacks are employed to learn a detector for distinguish-ing the perturbed examples, i.e., a) the embedding distancesbetween a perturbed query person image and its top-K re-trievals are generally larger than those between a benignquery image and its top-K retrievals, b) the embedding dis-tances among the top-K retrievals of a perturbed query im-age are larger than those of a benign query image, c) thetop-K retrievals of a benign query image obtained with mul-tiple expert ReID models tend to be consistent, which isnot preserved when attacks are present. Extensive exper-iments on the Market1501 and DukeMTMC-ReID datasetsshow that, as the first adversarial attack detection approachfor ReID,MEAADeffectively detects various adversarial at-tacks and achieves high ROC-AUC (over 97.5%).
