Entries by NEC Labs America

DeepTrack: Grouping RFID Tags Based on Spatio-temporal Proximity in Retail Spaces

/in /by

RFID applications for taking inventory and processing transactions in point-of-sale (POS) systems improve operational efficiency but are not designed to provide insights about customers’ interactions with products. We bridge this gap by solving the proximity grouping problem to identify groups of RFID tags that stay in close proximity to each other over time. We design DeepTrack, a framework that uses deep learning to automatically track the group of items carried by a customer during her shopping journey. This unearths hidden purchase behaviors helping retailers make better business decisions and paves the way for innovative shopping experiences such as seamless checkout (‘a la Amazon Go). DeepTrack employs a recurrent neural network (RNN) with the attention mechanism, to solve the proximity grouping problem in noisy settings without explicitly localizing tags. We tailor DeepTrack’s design to track not only mobile groups (products carried by customers) but also flexibly identify stationary tag groups (products on shelves). The key attribute of DeepTrack is that it only uses readily available tag data from commercial off-the-shelf RFID equipment. Our experiments demonstrate that, with only two hours training data, DeepTrack achieves a grouping accuracy of 98.18% (99.79%) when tracking eight mobile (stationary) groups.

Improving Disentangled Text Representation Learning with Information Theoretical Guidance

/in /by

Learning disentangled representations of natural language is essential for many NLP tasks, e.g., conditional text generation, style transfer, personalized dialogue systems, etc. Similar problems have been studied extensively for other forms of data, such as images and videos. However, the discrete nature of natural language makes the disentangling of textual representations more challenging (e.g., the manipulation over the data space cannot be easily achieved). Inspired by information theory, we propose a novel method that effectively manifests disentangled representations of text, without any supervision on semantics. A new mutual information upper bound is derived and leveraged to measure dependence between style and content. By minimizing this upper bound, the proposed method induces style and content embeddings into two independent low-dimensional spaces. Experiments on both conditional text generation and text-style transfer demonstrate the high quality of our disentangled representation in terms of content and style preservation.

On Optimal Multi-user Beam Alignment in Millimeter Wave Wireless Systems

/in /by

Directional transmission patterns (a.k.a. narrow beams) are the key to wireless communications in millimeter wave (mmWave) frequency bands which suffer from high path loss and severe shadowing. In addition, the propagation channel in mmWave frequencies incorporates only a few number of spatial clusters requiring a procedure to align the corresponding narrow beams with the angle of departure (AoD) of the channel clusters. The objective of this procedure, called beam alignment (BA) is to increase the beamforming gain for subsequent data communication. Several prior studies consider optimizing BA procedure to achieve various objectives such as reducing the BA overhead, increasing throughput, and reducing power consumption. While these studies mostly provide optimized BA schemes for scenarios with a single active user, there are often multiple active users in practical networks. Consequently, it is more efficient in terms of BA overhead and delay to design multi-user BA schemes which can perform beam management for multiple users collectively. This paper considers a class of multi-user BA schemes where the base station performs a one shot scan of the angular domain to simultaneously localize multiple users. The objective is to minimize the average of expected width of remaining uncertainty regions (UR) on the AoDs after receiving users’ feedbacks. Fundamental bounds on the optimal performance are analyzed using information theoretic tools. Furthermore, a BA optimization problem is formulated and a practical BA scheme, which provides significant gains compared to the beam sweeping used in 5G standard, is proposed.

Understanding Road Layout from Videos as a Whole

/in /by

In this paper, we address the problem of inferring the layout of complex road scenes from video sequences. To this end, we formulate it as a top-view road attributes prediction problem and our goal is to predict these attributes for each frame both accurately and consistently. In contrast to prior work, we exploit the following three novel aspects: leveraging camera motions in videos, including context cues and incorporating long-term video information. Specifically, we introduce a model that aims to enforce prediction consistency in videos. Our model consists of one LSTM and one Feature Transform Module (FTM). The former implicitly incorporates the consistency constraint with its hidden states, and the latter explicitly takes the camera motion into consideration when aggregating information along videos. Moreover, we propose to incorporate context information by introducing road participants, e.g. objects, into our model. When the entire video sequence is available, our model is also able to encode both local and global cues, e.g. information from both past and future frames. Experiments on two data sets show that: (1) Incorporating either global or contextual cues improves the prediction accuracy and leveraging both gives the best performance. (2) Introducing the LSTM and FTM modules improves the prediction consistency in videos. (3) The proposed method outperforms the SOTA by a large margin.

Towards Universal Representation Learning for Deep Face Recognition

/in /by

Recognizing wild faces is extremely hard as they appear with all kinds of variations. Traditional methods either train with specifically annotated variation data from target domains, or by introducing unlabeled target variation data to adapt from the training data. Instead, we propose a universal representation learning framework that can deal with larger variation unseen in the given training data without leveraging target domain knowledge. We firstly synthesize training data alongside some semantically meaningful variations, such as low resolution, occlusion and head pose. However, directly feeding the augmented data for training will not converge well as the newly introduced samples are mostly hard examples. We propose to split the feature embedding into multiple sub-embeddings, and associate different confidence values for each sub-embedding to smooth the training procedure. The sub-embeddings are further decorrelated by regularizing variation classification loss and variation adversarial loss on different partitions of them. Experiments show that our method achieves top performance on general face recognition datasets such as LFW and MegaFace, while significantly better on extreme benchmarks such as TinyFace and IJB-S.

Private-kNN Practical Differential Privacy for Computer Vision

/in /by

With increasing ethical and legal concerns on privacy for deep models in visual recognition, differential privacy has emerged as a mechanism to disguise membership of sensitive data in training datasets. Recent methods like Private Aggregation of Teacher Ensembles (PATE) leverage a large ensemble of teacher models trained on disjoint subsets of private data, to transfer knowledge to a student model with privacy guarantees. However, labeled vision data is often expensive and datasets, when split into many disjoint training sets, lead to significantly sub-optimal accuracy and thus hardly sustain good privacy bounds. We propose a practically data-efficient scheme based on private release of k-nearest neighbor (kNN) queries, which altogether avoids splitting the training dataset. Our approach allows the use of privacy-amplification by subsampling and iterative refinement of the kNN feature embedding. We rigorously analyze the theoretical properties of our method and demonstrate strong experimental performance on practical computer vision datasets for face attribute recognition and person reidentification. In particular, we achieve comparable or better accuracy than PATE while reducing more than 90% of the privacy loss, thereby providing the “most practical method to-date” for private deep learning in computer vision.

Peek-a-boo: Occlusion Reasoning in Indoor Scenes with Plane Representations

/in /by

We address the challenging task of occlusion-aware indoor 3D scene understanding. We represent scenes by a set of planes, where each one is defined by its normal, offset and two masks outlining (i) the extent of the visible part and (ii) the full region that consists of both visible and occluded parts of the plane. We infer these planes from a single input image with a novel neural network architecture. It consists of a two-branch category-specific module that aims to predict layout and objects of the scene separately so that different types of planes can be handled better. We also introduce a novel loss function based on plane warping that can leverage multiple views at training time for improved occlusion-aware reasoning. In order to train and evaluate our occlusion-reasoning model, we use the ScanNet dataset and propose (i) a strategy to automatically extract ground truth for both visible and hidden regions and (ii) a new evaluation metric that specifically focuses on the prediction in hidden regions. We empirically demonstrate that our proposed approach can achieve higher accuracy for occlusion reasoning compared to competitive baselines on the ScanNet dataset, e.g. 42.65% relative improvement on hidden regions.

S3VAE: Self-Supervised Sequential VAE for Representation Disentanglement and Data Generation

/in /by

We propose a sequential variational autoencoder to learn disentangled representations of sequential data (e.g., videos and audios) under self-supervision. Specifically, we exploit the benefits of some readily accessible supervision signals from input data itself or some off-the-shelf functional models and accordingly design auxiliary tasks for our model to utilize these signals. With the supervision of the signals, our model can easily disentangle the representation of an input sequence into static factors and dynamic factors (i.e., time-invariant and time-varying parts). Comprehensive experiments across videos and audios verify the effectiveness of our model on representation disentanglement and generation of sequential data, and demonstrate that, our model with self-supervision performs comparable to, if not better than, the fully-supervised model with ground truth labels, and outperforms state-of-the-art unsupervised models by a large margin.

15 Keypoints Is All You Need

/in /by

Pose-tracking is an important problem that requires identifying unique human pose-instances and matching them temporally across different frames in a video. However, existing pose-tracking methods are unable to accurately model temporal relationships and require significant computation, often computing the tracks offline. We present an efficient multi-person pose-tracking method, KeyTrack that only relies on keypoint information without using any RGB or optical flow to locate and track human keypoints in real-time. KeyTrack is a top-down approach that learns spatio-temporal pose relationships by modeling the multi-person pose-tracking problem as a novel Pose Entailment task using a Transformer-based architecture. Furthermore, KeyTrack uses a novel, parameter-free, keypoint refinement technique that improves the keypoint estimates used by the Transformers. We achieved state-of-the-art results on PoseTrack’17 and PoseTrack’18 benchmarks while using only a fraction of the computation used by most other methods for computing the tracking information.

At the Speed of Sound: Efficient Audio Scene Classification

/in /by

Efficient audio scene classification is essential for smart sensing platforms such as robots, medical monitoring, surveillance, or autonomous vehicles. We propose a retrieval-based scene classification architecture that combines recurrent neural networks and attention to compute embeddings for short audio segments. We train our framework using a custom audio loss function that captures both the relevance of audio segments within a scene and that of sound events within a segment. Using experiments on real audio scenes, we show that we can discriminate audio scenes with high accuracy after listening in for less than a second. This preserves 93% of the detection accuracy obtained after hearing the entire scene.