Search

We present a novel Automatic Target Recognition (ATR) system using open-vocabulary object detection and classification models. A primary advantage of this approach is that target classes can be defined just before runtime by a non-technical end user, using either a few natural language text descriptions of the target, or a few image exemplars, or both. Nuances in the desired targets can be expressed in natural language, which is useful for unique targets with little or no training data. We also implemented a novel combination of several techniques to improve performance, such as leveraging the additional information in the sequence of overlapping frames to perform tubelet identification (i.e., sequential bounding box matching), bounding box re-scoring, and tubelet linking. Additionally, we developed a technique to visualize the aggregate output of many overlapping frames as a mosaic of the area scanned during the aerial surveillance or reconnaissance, and a kernel density estimate (or heatmap) of the detected targets. We initially applied this ATR system to the use case of detecting and clearing unexploded ordinance on airfield runways and we are currently extending our research to other real-world applications., Comment: Accepted to the Thirty-Seventh Annual Conference on Innovative Applications of Artificial Intelligence (IAAI-25)

Despite the remarkable strides made by autoregressive language models, their potential is often hampered by the slow inference speeds inherent in sequential token generation. Blockwise parallel decoding (BPD) was proposed by Stern et al. as a method to improve inference speed of language models by simultaneously predicting multiple future tokens, termed block drafts, which are subsequently verified and conditionally accepted by the autoregressive model. This paper contributes to the understanding and improvement of block drafts in two ways. First, we analyze the token distributions produced by multiple prediction heads. Secondly, we leverage this analysis to develop algorithms to improve BPD inference speed by refining the block drafts using n-gram and neural language models. Experiments demonstrate that refined block drafts yield a +5-21% increase in block efficiency (i.e., the number of accepted tokens from the block draft) across diverse datasets.

Perception of the driving environment is critical for collision avoidance and route planning to ensure driving safety. Cooperative perception has been widely studied as an effective approach to addressing the shortcomings of single-vehicle perception. However, the practical limitations of vehicle-to-vehicle (V2V) communications have not been adequately investigated. In particular, current cooperative fusion models rely on supervised models and do not address dynamic performance degradation caused by arbitrary channel impairments. In this paper, a self-supervised adaptive weighting model is proposed for intermediate fusion to mitigate the adverse effects of channel distortion. The performance of cooperative perception is investigated in different system settings. Rician fading and imperfect channel state information (CSI) are also considered. Numerical results demonstrate that the proposed adaptive weighting algorithm significantly outperforms the benchmarks without weighting. Visualization examples validate that the proposed weighting algorithm can flexibly adapt to various channel conditions. Moreover, the adaptive weighting algorithm demonstrates good generalization to untrained channels and test datasets from different domains., Comment: accepted by IEEE Transactions on Intelligent Vehicles

Cooperative perception has been widely used in autonomous driving to alleviate the inherent limitation of single automated vehicle perception. To enable cooperation, vehicle-to-vehicle (V2V) communication plays an indispensable role. This work analyzes the performance of cooperative perception accounting for communications channel impairments. Different fusion methods and channel impairments are evaluated. A new late fusion scheme is proposed to leverage the robustness of intermediate features. In order to compress the data size incurred by cooperation, a convolution neural network-based autoencoder is adopted. Numerical results demonstrate that intermediate fusion is more robust to channel impairments than early fusion and late fusion, when the SNR is greater than 0 dB. Also, the proposed fusion scheme outperforms the conventional late fusion using detection outputs, and autoencoder provides a good compromise between detection accuracy and bandwidth usage.

In this work, we study the impact of Large-scale Language Models (LLM) on Automated Speech Recognition (ASR) of YouTube videos, which we use as a source for long-form ASR. We demonstrate up to 8\% relative reduction in Word Error Eate (WER) on US English (en-us) and code-switched Indian English (en-in) long-form ASR test sets and a reduction of up to 30\% relative on Salient Term Error Rate (STER) over a strong first-pass baseline that uses a maximum-entropy based language model. Improved lattice processing that results in a lattice with a proper (non-tree) digraph topology and carrying context from the 1-best hypothesis of the previous segment(s) results in significant wins in rescoring with LLMs. We also find that the gains in performance from the combination of LLMs trained on vast quantities of available data (such as C4) and conventional neural LMs is additive and significantly outperforms a strong first-pass baseline with a maximum entropy LM. Copyright 2023 IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works., Comment: 5 pages, accepted in ICASSP 2023

We introduce LAST, a LAttice-based Speech Transducer library in JAX. With an emphasis on flexibility, ease-of-use, and scalability, LAST implements differentiable weighted finite state automaton (WFSA) algorithms needed for training \& inference that scale to a large WFSA such as a recognition lattice over the entire utterance. Despite these WFSA algorithms being well-known in the literature, new challenges arise from performance characteristics of modern architectures, and from nuances in automatic differentiation. We describe a suite of generally applicable techniques employed in LAST to address these challenges, and demonstrate their effectiveness with benchmarks on TPUv3 and V100 GPU.

Existing training criteria in automatic speech recognition(ASR) permit the model to freely explore more than one time alignments between the feature and label sequences. In this paper, we use entropy to measure a model's uncertainty, i.e. how it chooses to distribute the probability mass over the set of allowed alignments. Furthermore, we evaluate the effect of entropy regularization in encouraging the model to distribute the probability mass only on a smaller subset of allowed alignments. Experiments show that entropy regularization enables a much simpler decoding method without sacrificing word error rate, and provides better time alignment quality.

We introduce a framework for adapting a virtual keyboard to individual user behavior by modifying a Gaussian spatial model to use personalized key center offset means and, optionally, learned covariances. Through numerous real-world studies, we determine the importance of training data quantity and weights, as well as the number of clusters into which to group keys to avoid overfitting. While past research has shown potential of this technique using artificially-simple virtual keyboards and games or fixed typing prompts, we demonstrate effectiveness using the highly-tuned Gboard app with a representative set of users and their real typing behaviors. Across a variety of top languages, we achieve small-but-significant improvements in both typing speed and decoder accuracy., Comment: 17 pages, to be published in the Proceedings of the 24th International Conference on Mobile Human-Computer Interaction (MobileHCI 2022)

We introduce the Globally Normalized Autoregressive Transducer (GNAT) for addressing the label bias problem in streaming speech recognition. Our solution admits a tractable exact computation of the denominator for the sequence-level normalization. Through theoretical and empirical results, we demonstrate that by switching to a globally normalized model, the word error rate gap between streaming and non-streaming speech-recognition models can be greatly reduced (by more than 50\% on the Librispeech dataset). This model is developed in a modular framework which encompasses all the common neural speech recognition models. The modularity of this framework enables controlled comparison of modelling choices and creation of new models.

Much of the literature on differential privacy focuses on item-level privacy, where loosely speaking, the goal is to provide privacy per item or training example. However, recently many practical applications such as federated learning require preserving privacy for all items of a single user, which is much harder to achieve. Therefore understanding the theoretical limit of user-level privacy becomes crucial. We study the fundamental problem of learning discrete distributions over $k$ symbols with user-level differential privacy. If each user has $m$ samples, we show that straightforward applications of Laplace or Gaussian mechanisms require the number of users to be $\mathcal{O}(k/(m\alpha^2) + k/\epsilon\alpha)$ to achieve an $\ell_1$ distance of $\alpha$ between the true and estimated distributions, with the privacy-induced penalty $k/\epsilon\alpha$ independent of the number of samples per user $m$. Moreover, we show that any mechanism that only operates on the final aggregate counts should require a user complexity of the same order. We then propose a mechanism such that the number of users scales as $\tilde{\mathcal{O}}(k/(m\alpha^2) + k/\sqrt{m}\epsilon\alpha)$ and hence the privacy penalty is $\tilde{\Theta}(\sqrt{m})$ times smaller compared to the standard mechanisms in certain settings of interest. We further show that the proposed mechanism is nearly-optimal under certain regimes. We also propose general techniques for obtaining lower bounds on restricted differentially private estimators and a lower bound on the total variation between binomial distributions, both of which might be of independent interest., Comment: NeurIPS 2020, 38 pages

This paper proposes and evaluates the hybrid autoregressive transducer (HAT) model, a time-synchronous encoderdecoder model that preserves the modularity of conventional automatic speech recognition systems. The HAT model provides a way to measure the quality of the internal language model that can be used to decide whether inference with an external language model is beneficial or not. This article also presents a finite context version of the HAT model that addresses the exposure bias problem and significantly simplifies the overall training and inference. We evaluate our proposed model on a large-scale voice search task. Our experiments show significant improvements in WER compared to the state-of-the-art approaches.

We propose algorithms to train production-quality n-gram language models using federated learning. Federated learning is a distributed computation platform that can be used to train global models for portable devices such as smart phones. Federated learning is especially relevant for applications handling privacy-sensitive data, such as virtual keyboards, because training is performed without the users' data ever leaving their devices. While the principles of federated learning are fairly generic, its methodology assumes that the underlying models are neural networks. However, virtual keyboards are typically powered by n-gram language models for latency reasons. We propose to train a recurrent neural network language model using the decentralized FederatedAveraging algorithm and to approximate this federated model server-side with an n-gram model that can be deployed to devices for fast inference. Our technical contributions include ways of handling large vocabularies, algorithms to correct capitalization errors in user data, and efficient finite state transducer algorithms to convert word language models to word-piece language models and vice versa. The n-gram language models trained with federated learning are compared to n-grams trained with traditional server-based algorithms using A/B tests on tens of millions of users of virtual keyboard. Results are presented for two languages, American English and Brazilian Portuguese. This work demonstrates that high-quality n-gram language models can be trained directly on client mobile devices without sensitive training data ever leaving the devices., Comment: 10 pages

Weighted finite automata (WFA) are often used to represent probabilistic models, such as $n$-gram language models, since they are efficient for recognition tasks in time and space. The probabilistic source to be represented as a WFA, however, may come in many forms. Given a generic probabilistic model over sequences, we propose an algorithm to approximate it as a weighted finite automaton such that the Kullback-Leiber divergence between the source model and the WFA target model is minimized. The proposed algorithm involves a counting step and a difference of convex optimization step, both of which can be performed efficiently. We demonstrate the usefulness of our approach on various tasks, including distilling $n$-gram models from neural models, building compact language models, and building open-vocabulary character models. The algorithms used for these experiments are available in an open-source software library.

We propose a finite-state transducer (FST) representation for the models used to decode keyboard inputs on mobile devices. Drawing from learnings from the field of speech recognition, we describe a decoding framework that can satisfy the strict memory and latency constraints of keyboard input. We extend this framework to support functionalities typically not present in speech recognition, such as literal decoding, autocorrections, word completions, and next word predictions. We describe the general framework of what we call for short the keyboard "FST decoder" as well as the implementation details that are new compared to a speech FST decoder. We demonstrate that the FST decoder enables new UX features such as post-corrections. Finally, we sketch how this decoder can support advanced features such as personalization and contextualization.