Search results

This article presents a whisper speech detector in the far-field domain. The proposed system consists of a long short-term memory (LSTM) neural network trained on log-filterbank energy (LFBE) acoustic features. This model is trained and evaluated on recordings of human interactions with voice-controlled, far-field devices in whisper and normal phonation modes. We compare multiple inference approaches for

The state of the art in probabilistic demand forecasting [40] minimizes Quantile Loss to predict the future demand quantiles for different horizons. However, since quantiles aren’t additive, in order to predict the total demand for any wider future interval all required intervals are usually appended to the target vector during model training. The separate optimization of these overlapping intervals can

The web contains countless semi-structured websites, which can be a rich source of information for populating knowledge bases. Existing methods for extracting relations from the DOM trees of semi-structured webpages can achieve high precision and recall only when manual annotations for each website are available. Although there have been efforts to learn extractors from automatically generated labels, these

Extreme multi-label classification (XMC) aims at assigning to an instance the most relevant subset of labels from a colossal label set. There have been some success in formulating the multi-label problem as sequence-to-sequence (Seq2Seq) learning, where the positive class labels of each input instance are used as the corresponding output sequence. Seq2Seq methods, nonetheless, have not yet been scalable

Training large neural networks requires distributing learning across multiple workers, where the cost of communicating gradients can be a significant bottleneck. SIGNSGD alleviates this problem by transmitting just the sign of each minibatch stochastic gradient. We prove that it can get the best of both worlds: compressed gradients and SGD-level convergence rate. The relative `1/`2 geometry of gradients

When learning POS taggers and syntactic chunkers for low-resource languages, different resources may be available, and often all we have is a small tag dictionary, motivating type-constrained unsupervised induction. Even small dictionaries can improve the performance of unsupervised induction algorithms. This paper shows that performance can be further improved by including data that is readily available

Large scale Natural Language Understanding (NLU) systems are typically trained on large quantities of data, requiring a fast and scalable training strategy. A typical design for NLU systems consists of domain-level NLU modules (domain classifier, intent classifier and named entity recognizer). Hypotheses (NLU interpretations consisting of various intent+slot combinations) from these domain specific modules

Goal-oriented dialogue systems typically rely on components specifically developed for a single task or domain. This limits such systems in two different ways: If there is an update in the task domain, the dialogue system usually needs to be updated or completely re-trained. It is also harder to extend such dialogue systems to different and multiple domains. The dialogue state tracker in conventional dialogue

The Amazon Echo is a hands-free smart home speaker you control with your voice. The first important step in enabling a delightful customer experience with an Echo or other Alexa-enabled device is wake word detection, so accurate detection of “Alexa” or substitute wake words is critical. It is challenging to build a wake word system with low error rates when there are limited computation resources on the device and it's in the presence of background noise such as speech or music.

Our team has been developed from a deep learning study group at SNU.

We present a platform built on large-scale, data-centric machine learning (ML) approaches, whose particular focus is demand forecasting in retail. At its core, this platform enables the training and application of probabilistic demand forecasting models, and provides convenient abstractions and support functionality for forecasting problems. The platform comprises of a complex end-to-end machine learning

We present a novel, large-scale ensemble-based system for the Amazon Alexa Prize competition. Our system leverages state-of-the-art methods from deep learning and reinforcement learning. We carry out A/B testing experiments with real-world users and demonstrate that our approach yields substantial improvements over several baseline systems. During the competition semi-finals, our best performing system

We describe our 2018 Alexa prize system (called ‘Alana’) which consists of an ensemble of bots, combining rule-based and machine learning systems. This paper reports on the version of the system developed and evaluated in the semifinals of the 2018 competition (i.e. up to 15 August 2018), but not on subsequent enhancements. The main advances over our 2017 Alana system are: (1) a deeper Natural Language

We present BYU-EVE, an open domain dialogue architecture that combines the strengths of hand-crafted rules, deep learning, and structured knowledge graph traversal in order to create satisfying user experiences. Rather than viewing dialogue as a strict mapping between input and output texts, EVE treats conversations as a collaborative process in which two jointly coordinating agents chart a trajectory through

In this paper we present Fantom, a social chatbot competing in the Amazon Alexa Prize 2018. The system uses a dialog graph for retrieving an approximation of the current dialog context in order to find suitable response candidates in this context. The graph is gradually built using user utterances from actual interactions, and system responses suggested by crowd workers. To this end, we developed an automatic

Recommendation systems have developed beyond simple matrix factorization to focus on two important sources of information: the temporal order of events (Hidasi et al., 2015) and side (e.g., spatial) information encoded in user and item features (Rendle, 2012). However, state-of-art temporal modeling is often limited by model capacity for long user histories. In addition, meta data are rarely used in generic

Sounding Board is a conversational agent that engages users through thoughtful and informative discussion.

Amazon Alexa currently has more than 40,000 third-party skills, which customers use to get information, perform tasks, play games, and more. To make it easier for customers to find and engage with skills, we are moving toward skill invocation that doesn’t require mentioning a skill by name (as highlighted in a recent post).

In recent years, the amount of textual information produced daily has increased exponentially. This information explosion has been accelerated by the ease with which data can be shared across the web. Most of the textual information is generated as free-form text, and only a small fraction is available in structured format (Wikidata, Freebase etc.) that can be processed and analyzed directly by machines.

Picture walking into your house after a long day. As you step in, your favorite music starts playing in the background and the lights dim to suit the weather outside.