Search results

Traditionally, Alexa has interpreted customer requests according to their intents and slots. If you say, “Alexa, play ‘What’s Going On?’ by Marvin Gaye,” the intent should be PlayMusic, and “‘What’s Going On?’” and “Marvin Gaye” should fill the slots SongName and ArtistName.

Large language models (LLMs) have been widely deployed and have achieved remarkable success in downstream tasks. However, their high latency continues to pose challenges for real-time applications that require fast inference, and the need to train and deploy distinct models for different hardware constraints increases both financial and computational costs. To address this, we propose Nested Matrix Learning

Enterprise environments differ fundamentally from the clean settings assumed in LLM research: knowledge is distributed across heterogeneous sources, often incomplete or inconsistent, and key procedural logic is implicitly encoded in artifacts rather than explicitly documented. In such settings, retrieval-based approaches are insufficient, as no single source contains the full workflow. We propose a replication-driven

Amazon Research Award recipient

Deep Metric Learning (DML) methods aim at learning an embedding space in which distances are closely related to the inherent semantic similarity of the inputs. Previous studies have shown that popular benchmark datasets often contain numerous wrong labels, and DML methods are susceptible to them. Intending to study the effect of realistic noise, we create an ontology of the classes in a dataset and use

Data augmentation is an inexpensive way to increase training data diversity, and is commonly achieved via transformations of existing data. For tasks such as classification, there is a good case for learning representations of the data that are invariant to such transformations, yet this is not explicitly enforced by classification losses such as the cross-entropy loss. This paper investigates the use of

Accelerated magnetic resonance imaging resorts to either Fourier-domain subsampling or better reconstruction algorithms to deal with fewer measurements while still generating medical images of high quality. Determining the optimal sampling strategy given a fixed reconstruction protocol often has combinatorial complexity. In this work, we apply double deep Q-learning and REINFORCE algorithms to learn the

KingFisher is a team of several graduate and undergraduate students from the University of Illinois, Urbana-Champaign led by faculty advisor Julia Hockenmaier.

We present ProbHardE2E, a probabilistic forecasting framework that incorporates hard operational/physical constraints, and provides uncertainty quantification. Our methodology uses a novel differentiable probabilistic projection layer (DPPL) that can be combined with a wide range of neural network architectures. DPPL allows the model to learn the system in an end-to-end manner, compared to other approaches

In industrial settings, it is often necessary to achieve language-level accuracy targets. For example, Amazon business teams need to build multilingual product classifiers that operate accurately in all European languages. It is unacceptable for the accuracy of product classification to meet the target in one language (e.g, English), while falling below the target in other languages (e.g, Portuguese). To

We show that the ability of a neural network to integrate information from diverse sources hinges critically on being exposed to properly correlated signals during the early phases of training. Interfering with the learning process during this initial stage can permanently impair the development of a skill, both in artificial and biological systems where the phenomenon is known as a critical learning period

In deep metric learning for visual recognition, the calibration of distance thresholds is crucial for achieving desired model performance in the true positive rates (TPR) or true negative rates (TNR). However, calibrating this threshold presents challenges in open-world scenarios, where the test classes can be entirely disjoint from those encountered during training. We define the problem of finding distance

Humans generally use natural language (NL) to communicate task requirements to each other. Ideally, NL should also be usable for communicating goals to autonomous machines (e.g., robots) to minimize friction in task specification. However, understanding and mapping NL goals to sequences of states and actions is challenging. Specifically, existing work along these lines has encountered difficulty in generalizing

We consider the task of predicting subjective fashion traits from images. Specifically, we are interested in understanding which outfit actually better suites the user. Since these traits are highly subjective, they tend to be noisier. One solution is to annotate each example several times, but this makes it hard to collect large amounts of data.

In this work, we demonstrate that due to the inadequacies in the existing evaluation protocols and datasets, there is a need to revisit and comprehensively examine the multimodal Zero-Shot Learning (MZSL) problem formulation. Specifically, we address two major challenges faced by current MZSL approaches; (1) Established baselines are frequently incomparable and occasionally even flawed since existing evaluation

Large language models (LLMs) have demonstrated impressive performance on a number of natural language processing tasks, such as question answering and text summarization. However, their performance on sequence labeling tasks, such as intent classification and slot filling (IC-SF), which is a central component in personal assistant systems, lags significantly behind discriminative models. Furthermore, there

Pitch detection is a fundamental problem in speech processing as F0 is used in a large number of applications. Recent papers have proposed deep learning for robust pitch tracking. In this letter, we consider voicing detection as a classification problem and F0 contour estimation as a regression problem. For both tasks, acoustic features from multiple domains and traditional machine learning methods are

Speech emotion recognition (SER) is a key technology to enable more natural human-machine communication. However, SER has long suffered from a lack of public large-scale labeled datasets. To circumvent this problem, we investigate how unsupervised representation learning on unlabeled datasets can benefit SER. We show that the contrastive predictive coding (CPC) method can learn salient representations from

We introduce Robust Training with Trust Scores (RT2S), a framework to train machine learning classifiers with potentially noisy labels. RT2S calculates a trust score for each training sample, which indicates the quality of its corresponding label. These trust scores are employed as sample weights during training and optionally during threshold optimization. The trust scores are generated from two sources

Rapid growth of deep learning models in recent years for robot and fraud detection has led to significant improvement in precision and recall but has also created a challenge for explainability and trust in the model decisions. In this paper, we propose a scalable multitiered framework that generates explainable network request level signatures for crawler bots on a large e-commerce advertising program.