arXiv subjects

Steve Young

Publications and source records attributed to Steve Young.

At least 19 recordsLinked to original sources

arXiv2025

Temporal Coarse Graining for Classical Stochastic Noise in Quantum Systems

Simulations of quantum systems with Hamiltonian classical stochastic noise can be challenging when the noise exhibits temporal correlations over a multitude of time scales, such as for $1/f$ noise in solid-state quantum information processors. Here we present an approach for simulating Hamiltonian classical stochastic noise that performs temporal coarse-graining by effectively integrating out the high-frequency components of the noise. We focus on the case where the stochastic noise can be expressed as a sum of Ornstein-Uhlenbeck processes. Temporal coarse-graining is then achieved by conditioning the stochastic process on a coarse realization of the noise, expressing the conditioned stochastic process in terms of a sum of smooth, deterministic functions and bridge processes with boundaries fixed at zero, and performing the ensemble average over the bridge processes. For Ornstein-Uhlenbeck processes, the deterministic components capture all dependence on the coarse realization, and the stochastic bridge processes are not only independent but taken from the same distribution with correlators that can be expressed analytically, allowing the associated noise propagators to be precomputed once for all simulations. This combination of noise trajectories on a coarse time grid and ensemble averaging over bridge processes has practical advantages, such as a simple concatenation rule, that we highlight with numerical examples.

quant-ph↗

arXiv2024

Model validation and error attribution for a drifting qubit

Qubit performance is often reported in terms of a variety of single-value metrics, each providing a facet of the underlying noise mechanism limiting performance. However, the value of these metrics may drift over long time-scales, and reporting a single number for qubit performance fails to account for the low-frequency noise processes that give rise to this drift. In this work, we demonstrate how we can use the distribution of these values to validate or invalidate candidate noise models. We focus on the case of randomized benchmarking (RB), where typically a single error rate is reported but this error rate can drift over time when multiple passes of RB are performed. We show that using a statistical test as simple as the Kolmogorov-Smirnov statistic on the distribution of RB error rates can be used to rule out noise models, assuming the experiment is performed over a long enough time interval to capture relevant low frequency noise. With confidence in a noise model, we show how care must be exercised when performing error attribution using the distribution of drifting RB error rate.

quant-ph↗

arXiv2021

Dielectric Resonator Antenna Coupled Antimonide-Based Detectors (DRACAD) For the Infrared

In an infrared photodetector, noise current (dark current) is generated throughout the volume of the detector. Reducing the volume will reduce dark current, but the corresponding smaller area will also reduce the received signal. By using a separate antenna to receive light, one can reduce the detector area without reducing the signal, thereby increasing its signal to noise ratio (SNR). Here, we present a dielectric resonator antenna (DRA) coupled infrared photodetector. Its performance is compared to a conventional resonant cavity enhanced slab detector. The Noise Equivalent Power (NEP) is used as a figure of merit for the comparison. Formulas for the NEP are derived for both cases. A pBp photodiode detector is assumed in the comparison. The active region of the photodiode is an InAs/GaSb Type II Superlattice (T2SL). A Genetic Algorithm is used to optimize the dimensions of the detector and the DRA to achieve the smallest NEP. The result is a photodetector that converts over 85% of the incident light into carriers with a volume reduced by 95%. This optimal geometry leads to a NEP reduced by 6.02dB over that of the conventional resonant cavity enhanced slab detector.

physics.ins-det↗

arXiv2019

Unsupervised Inflection Generation Using Neural Language Modeling

The use of Deep Neural Network architectures for Language Modeling has recently seen a tremendous increase in interest in the field of NLP with the advent of transfer learning and the shift in focus from rule-based and predictive models (supervised learning) to generative or unsupervised models to solve the long-standing problems in NLP like Information Extraction or Question Answering. While this shift has worked greatly for languages lacking in inflectional morphology, such as English, challenges still arise when trying to build similar systems for morphologically-rich languages, since their individual words shift forms in context more often. In this paper we investigate the extent to which these new unsupervised or generative techniques can serve to alleviate the type-token ratio disparity in morphologically rich languages. We apply an off-the-shelf neural language modeling library to the newly introduced task of unsupervised inflection generation in the nominal domain of three morphologically rich languages: Romanian, German, and Finnish. We show that this neural language model architecture can successfully generate the full inflection table of nouns without needing any pre-training on large, wikipedia-sized corpora, as long as the model is shown enough inflection examples. In fact, our experiments show that pre-training hinders the generation performance.

cs.CL↗

arXiv2019

Addressing Objects and Their Relations: The Conversational Entity Dialogue Model

Statistical spoken dialogue systems usually rely on a single- or multi-domain dialogue model that is restricted in its capabilities of modelling complex dialogue structures, e.g., relations. In this work, we propose a novel dialogue model that is centred around entities and is able to model relations as well as multiple entities of the same type. We demonstrate in a prototype implementation benefits of relation modelling on the dialogue level and show that a trained policy using these relations outperforms the multi-domain baseline. Furthermore, we show that by modelling the relations on the dialogue level, the system is capable of processing relations present in the user input and even learns to address them in the system response.

cs.CL↗

arXiv2018

Nearly Zero-Shot Learning for Semantic Decoding in Spoken Dialogue Systems

This paper presents two ways of dealing with scarce data in semantic decoding using N-Best speech recognition hypotheses. First, we learn features by using a deep learning architecture in which the weights for the unknown and known categories are jointly optimised. Second, an unsupervised method is used for further tuning the weights. Sharing weights injects prior knowledge to unknown categories. The unsupervised tuning (i.e. the risk minimisation) improves the F-Measure when recognising nearly zero-shot data on the DSTC3 corpus. This unsupervised method can be applied subject to two assumptions: the rank of the class marginal is assumed to be known and the class-conditional scores of the classifier are assumed to follow a Gaussian distribution.

cs.CL↗

arXiv2018

A Benchmarking Environment for Reinforcement Learning Based Task Oriented Dialogue Management

Dialogue assistants are rapidly becoming an indispensable daily aid. To avoid the significant effort needed to hand-craft the required dialogue flow, the Dialogue Management (DM) module can be cast as a continuous Markov Decision Process (MDP) and trained through Reinforcement Learning (RL). Several RL models have been investigated over recent years. However, the lack of a common benchmarking framework makes it difficult to perform a fair comparison between different models and their capability to generalise to different environments. Therefore, this paper proposes a set of challenging simulated environments for dialogue model development and evaluation. To provide some baselines, we investigate a number of representative parametric algorithms, namely deep reinforcement learning algorithms - DQN, A2C and Natural Actor-Critic and compare them to a non-parametric model, GP-SARSA. Both the environments and policy models are implemented using the publicly available PyDial toolkit and released on-line, in order to establish a testbed framework for further experiments and to facilitate experimental reproducibility.

stat.ML↗

arXiv2017

Holographic constraints on Bjorken hydrodynamics at finite coupling

In large-$N_c$ conformal field theories with classical holographic duals, inverse coupling constant corrections are obtained by considering higher-derivative terms in the corresponding gravity theory. In this work, we use type IIB supergravity and bottom-up Gauss-Bonnet gravity to study the dynamics of boost-invariant Bjorken hydrodynamics at finite coupling. We analyze the time-dependent decay properties of non-local observables (scalar two-point functions and Wilson loops) probing the different models of Bjorken flow and show that they can be expressed generically in terms of a few field theory parameters. In addition, our computations provide an analytically quantifiable probe of the coupling-dependent validity of hydrodynamics at early times in a simple model of heavy-ion collisions, which is an observable closely analogous to the hydrodynamization time of a quark-gluon plasma. We find that to third order in the hydrodynamic expansion, the convergence of hydrodynamics is improved and that generically, as expected from field theory considerations and recent holographic results, the applicability of hydrodynamics is delayed as the field theory coupling decreases.

hep-th↗

arXiv2017

Reward-Balancing for Statistical Spoken Dialogue Systems using Multi-objective Reinforcement Learning

Reinforcement learning is widely used for dialogue policy optimization where the reward function often consists of more than one component, e.g., the dialogue success and the dialogue length. In this work, we propose a structured method for finding a good balance between these components by searching for the optimal reward component weighting. To render this search feasible, we use multi-objective reinforcement learning to significantly reduce the number of training dialogues required. We apply our proposed method to find optimized component weights for six domains and compare them to a default baseline.

cs.CL↗

arXiv2017

Sample-efficient Actor-Critic Reinforcement Learning with Supervised Data for Dialogue Management

Deep reinforcement learning (RL) methods have significant potential for dialogue policy optimisation. However, they suffer from a poor performance in the early stages of learning. This is especially problematic for on-line learning with real users. Two approaches are introduced to tackle this problem. Firstly, to speed up the learning process, two sample-efficient neural networks algorithms: trust region actor-critic with experience replay (TRACER) and episodic natural actor-critic with experience replay (eNACER) are presented. For TRACER, the trust region helps to control the learning step size and avoid catastrophic model changes. For eNACER, the natural gradient identifies the steepest ascent direction in policy space to speed up the convergence. Both models employ off-policy learning with experience replay to improve sample-efficiency. Secondly, to mitigate the cold start issue, a corpus of demonstration data is utilised to pre-train the models prior to on-line reinforcement learning. Combining these two approaches, we demonstrate a practical approach to learn deep RL-based dialogue policies and demonstrate their effectiveness in a task-oriented information seeking domain.

cs.CL↗

arXiv2017

Semantic Specialisation of Distributional Word Vector Spaces using Monolingual and Cross-Lingual Constraints

We present Attract-Repel, an algorithm for improving the semantic quality of word vectors by injecting constraints extracted from lexical resources. Attract-Repel facilitates the use of constraints from mono- and cross-lingual resources, yielding semantically specialised cross-lingual vector spaces. Our evaluation shows that the method can make use of existing cross-lingual lexicons to construct high-quality vector spaces for a plethora of different languages, facilitating semantic transfer from high- to lower-resource ones. The effectiveness of our approach is demonstrated with state-of-the-art results on semantic similarity datasets in six languages. We next show that Attract-Repel-specialised vectors boost performance in the downstream task of dialogue state tracking (DST) across multiple languages. Finally, we show that cross-lingual vector spaces produced by our algorithm facilitate the training of multilingual DST models, which brings further performance improvements.

cs.CL↗

arXiv2017

Morph-fitting: Fine-Tuning Word Vector Spaces with Simple Language-Specific Rules

Morphologically rich languages accentuate two properties of distributional vector space models: 1) the difficulty of inducing accurate representations for low-frequency word forms; and 2) insensitivity to distinct lexical relations that have similar distributional signatures. These effects are detrimental for language understanding systems, which may infer that 'inexpensive' is a rephrasing for 'expensive' or may not associate 'acquire' with 'acquires'. In this work, we propose a novel morph-fitting procedure which moves past the use of curated semantic lexicons for improving distributional vector spaces. Instead, our method injects morphological constraints generated using simple language-specific rules, pulling inflectional forms of the same word close together and pushing derivational antonyms far apart. In intrinsic evaluation over four languages, we show that our approach: 1) improves low-frequency word estimates; and 2) boosts the semantic quality of the entire word vector collection. Finally, we show that morph-fitted vectors yield large gains in the downstream task of dialogue state tracking, highlighting the importance of morphology for tackling long-tail phenomena in language understanding tasks.

cs.CL↗

arXiv2017

Latent Intention Dialogue Models

Developing a dialogue agent that is capable of making autonomous decisions and communicating by natural language is one of the long-term goals of machine learning research. Traditional approaches either rely on hand-crafting a small state-action set for applying reinforcement learning that is not scalable or constructing deterministic models for learning dialogue sentences that fail to capture natural conversational variability. In this paper, we propose a Latent Intention Dialogue Model (LIDM) that employs a discrete latent variable to learn underlying dialogue intentions in the framework of neural variational inference. In a goal-oriented dialogue scenario, these latent intentions can be interpreted as actions guiding the generation of machine responses, which can be further refined autonomously by reinforcement learning. The experimental evaluation of LIDM shows that the model out-performs published benchmarks for both corpus-based and human evaluation, demonstrating the effectiveness of discrete latent variable models for learning goal-oriented dialogues.

cs.CL↗

arXiv2017

A Network-based End-to-End Trainable Task-oriented Dialogue System

Teaching machines to accomplish tasks by conversing naturally with humans is challenging. Currently, developing task-oriented dialogue systems requires creating multiple components and typically this involves either a large amount of handcrafting, or acquiring costly labelled datasets to solve a statistical learning problem for each component. In this work we introduce a neural network-based text-in, text-out end-to-end trainable goal-oriented dialogue system along with a new way of collecting dialogue data based on a novel pipe-lined Wizard-of-Oz framework. This approach allows us to develop dialogue systems easily and without making too many assumptions about the task at hand. The results show that the model can converse with human subjects naturally whilst helping them to accomplish tasks in a restaurant search domain.

cs.CL↗

arXiv2017

Neural Belief Tracker: Data-Driven Dialogue State Tracking

One of the core components of modern spoken dialogue systems is the belief tracker, which estimates the user's goal at every step of the dialogue. However, most current approaches have difficulty scaling to larger, more complex dialogue domains. This is due to their dependency on either: a) Spoken Language Understanding models that require large amounts of annotated training data; or b) hand-crafted lexicons for capturing some of the linguistic variation in users' language. We propose a novel Neural Belief Tracking (NBT) framework which overcomes these problems by building on recent advances in representation learning. NBT models reason over pre-trained word vectors, learning to compose them into distributed representations of user utterances and dialogue context. Our evaluation on two datasets shows that this approach surpasses past limitations, matching the performance of state-of-the-art models which rely on hand-crafted semantic lexicons and outperforming them when such lexicons are not provided.

cs.CL↗

arXiv2016

Exploiting Sentence and Context Representations in Deep Neural Models for Spoken Language Understanding

This paper presents a deep learning architecture for the semantic decoder component of a Statistical Spoken Dialogue System. In a slot-filling dialogue, the semantic decoder predicts the dialogue act and a set of slot-value pairs from a set of n-best hypotheses returned by the Automatic Speech Recognition. Most current models for spoken language understanding assume (i) word-aligned semantic annotations as in sequence taggers and (ii) delexicalisation, or a mapping of input words to domain-specific concepts using heuristics that try to capture morphological variation but that do not scale to other domains nor to language variation (e.g., morphology, synonyms, paraphrasing ). In this work the semantic decoder is trained using unaligned semantic annotations and it uses distributed semantic representation learning to overcome the limitations of explicit delexicalisation. The proposed architecture uses a convolutional neural network for the sentence representation and a long-short term memory network for the context representation. Results are presented for the publicly available DSTC2 corpus and an In-car corpus which is similar to DSTC2 but has a significantly higher word error rate (WER).

cs.AI↗

arXiv2016

Dialogue manager domain adaptation using Gaussian process reinforcement learning

Spoken dialogue systems allow humans to interact with machines using natural speech. As such, they have many benefits. By using speech as the primary communication medium, a computer interface can facilitate swift, human-like acquisition of information. In recent years, speech interfaces have become ever more popular, as is evident from the rise of personal assistants such as Siri, Google Now, Cortana and Amazon Alexa. Recently, data-driven machine learning methods have been applied to dialogue modelling and the results achieved for limited-domain applications are comparable to or outperform traditional approaches. Methods based on Gaussian processes are particularly effective as they enable good models to be estimated from limited training data. Furthermore, they provide an explicit estimate of the uncertainty which is particularly useful for reinforcement learning. This article explores the additional steps that are necessary to extend these methods to model multiple dialogue domains. We show that Gaussian process reinforcement learning is an elegant framework that naturally supports a range of methods, including prior knowledge, Bayesian committee machines and multi-agent learning, for facilitating extensible and adaptable dialogue systems.

cs.CL↗

arXiv2016

Conditional Generation and Snapshot Learning in Neural Dialogue Systems

Recently a variety of LSTM-based conditional language models (LM) have been applied across a range of language generation tasks. In this work we study various model architectures and different ways to represent and aggregate the source information in an end-to-end neural dialogue system framework. A method called snapshot learning is also proposed to facilitate learning from supervised sequential signals by applying a companion cross-entropy objective function to the conditioning vector. The experimental and analytical results demonstrate firstly that competition occurs between the conditioning vector and the LM, and the differing architectures provide different trade-offs between the two. Secondly, the discriminative power and transparency of the conditioning vector is key to providing both model interpretability and better performance. Thirdly, snapshot learning leads to consistent performance improvements independent of which architecture is used.

cs.CL↗