arXiv subjects

Jimmy Lin

Publications and source records attributed to Jimmy Lin.

At least 217 records · Page 12Linked to original sources

arXiv2019

Simple BERT Models for Relation Extraction and Semantic Role Labeling

We present simple BERT-based models for relation extraction and semantic role labeling. In recent years, state-of-the-art performance has been achieved using neural models by incorporating lexical and syntactic features such as part-of-speech tags and dependency trees. In this paper, extensive experiments on datasets for these two tasks show that without using any external features, a simple BERT-based model can achieve state-of-the-art performance. To our knowledge, we are the first to successfully apply BERT in this manner. Our models provide strong baselines for future research.

cs.CL↗

arXiv2019

Distilling Task-Specific Knowledge from BERT into Simple Neural Networks

In the natural language processing literature, neural networks are becoming increasingly deeper and complex. The recent poster child of this trend is the deep language representation model, which includes BERT, ELMo, and GPT. These developments have led to the conviction that previous-generation, shallower neural networks for language understanding are obsolete. In this paper, however, we demonstrate that rudimentary, lightweight neural networks can still be made competitive without architecture changes, external training data, or additional input features. We propose to distill knowledge from BERT, a state-of-the-art language representation model, into a single-layer BiLSTM, as well as its siamese counterpart for sentence-pair tasks. Across multiple datasets in paraphrasing, natural language inference, and sentiment classification, we achieve comparable results with ELMo, while using roughly 100 times fewer parameters and 15 times less inference time.

cs.CL↗

arXiv2019

Simple Applications of BERT for Ad Hoc Document Retrieval

Following recent successes in applying BERT to question answering, we explore simple applications to ad hoc document retrieval. This required confronting the challenge posed by documents that are typically longer than the length of input BERT was designed to handle. We address this issue by applying inference on sentences individually, and then aggregating sentence scores to produce document scores. Experiments on TREC microblog and newswire test collections show that our approach is simple yet effective, as we report the highest average precision on these datasets by neural approaches that we are aware of.

cs.IR↗

arXiv2019

Matching Entities Across Different Knowledge Graphs with Graph Embeddings

This paper explores the problem of matching entities across different knowledge graphs. Given a query entity in one knowledge graph, we wish to find the corresponding real-world entity in another knowledge graph. We formalize this problem and present two large-scale datasets for this task based on exiting cross-ontology links between DBpedia and Wikidata, focused on several hundred thousand ambiguous entities. Using a classification-based approach, we find that a simple multi-layered perceptron based on representations derived from RDF2Vec graph embeddings of entities in each knowledge graph is sufficient to achieve high accuracy, with only small amounts of training data. The contributions of our work are datasets for examining this problem and strong baselines on which future work can be based.

cs.CL↗

arXiv2018

Streaming Voice Query Recognition using Causal Convolutional Recurrent Neural Networks

Voice-enabled commercial products are ubiquitous, typically enabled by lightweight on-device keyword spotting (KWS) and full automatic speech recognition (ASR) in the cloud. ASR systems require significant computational resources in training and for inference, not to mention copious amounts of annotated speech data. KWS systems, on the other hand, are less resource-intensive but have limited capabilities. On the Comcast Xfinity X1 entertainment platform, we explore a middle ground between ASR and KWS: We introduce a novel, resource-efficient neural network for voice query recognition that is much more accurate than state-of-the-art CNNs for KWS, yet can be easily trained and deployed with limited resources. On an evaluation dataset representing the top 200 voice queries, we achieve a low false alarm rate of 1% and a query error rate of 6%. Our model performs inference 8.24x faster than the current ASR system.

cs.CL↗

arXiv2018

FLOPs as a Direct Optimization Objective for Learning Sparse Neural Networks

There exists a plethora of techniques for inducing structured sparsity in parametric models during the optimization process, with the final goal of resource-efficient inference. However, few methods target a specific number of floating-point operations (FLOPs) as part of the optimization objective, despite many reporting FLOPs as part of the results. Furthermore, a one-size-fits-all approach ignores realistic system constraints, which differ significantly between, say, a GPU and a mobile phone -- FLOPs on the former incur less latency than on the latter; thus, it is important for practitioners to be able to specify a target number of FLOPs during model compression. In this work, we extend a state-of-the-art technique to directly incorporate FLOPs as part of the optimization objective and show that, given a desired FLOPs requirement, different neural networks can be successfully trained for image classification.

cs.LG↗

arXiv2018

Progress and Tradeoffs in Neural Language Models

In recent years, we have witnessed a dramatic shift towards techniques driven by neural networks for a variety of NLP tasks. Undoubtedly, neural language models (NLMs) have reduced perplexity by impressive amounts. This progress, however, comes at a substantial cost in performance, in terms of inference latency and energy consumption, which is particularly of concern in deployments on mobile devices. This paper, which examines the quality-performance tradeoff of various language modeling techniques, represents to our knowledge the first to make this observation. We compare state-of-the-art NLMs with "classic" Kneser-Ney (KN) LMs in terms of energy usage, latency, perplexity, and prediction accuracy using two standard benchmarks. On a Raspberry Pi, we find that orders of increase in latency and energy usage correspond to less change in perplexity, while the difference is much less pronounced on a desktop.

cs.CL↗

arXiv2018

JavaScript Convolutional Neural Networks for Keyword Spotting in the Browser: An Experimental Analysis

Used for simple commands recognition on devices from smart routers to mobile phones, keyword spotting systems are everywhere. Ubiquitous as well are web applications, which have grown in popularity and complexity over the last decade with significant improvements in usability under cross-platform conditions. However, despite their obvious advantage in natural language interaction, voice-enabled web applications are still far and few between. In this work, we attempt to bridge this gap by bringing keyword spotting capabilities directly into the browser. To our knowledge, we are the first to demonstrate a fully-functional implementation of convolutional neural networks in pure JavaScript that runs in any standards-compliant browser. We also apply network slimming, a model compression technique, to explore the accuracy-efficiency tradeoffs, reporting latency measurements on a range of devices and software. Overall, our robust, cross-device implementation for keyword spotting realizes a new paradigm for serving neural network applications, and one of our slim models reduces latency by 66% with a minimal decrease in accuracy of 4% from 94% to 90%.

cs.CY↗

arXiv2018

Serverless Data Analytics with Flint

Serverless architectures organized around loosely-coupled function invocations represent an emerging design for many applications. Recent work mostly focuses on user-facing products and event-driven processing pipelines. In this paper, we explore a completely different part of the application space and examine the feasibility of analytical processing on big data using a serverless architecture. We present Flint, a prototype Spark execution engine that takes advantage of AWS Lambda to provide a pure pay-as-you-go cost model. With Flint, a developer uses PySpark exactly as before, but without needing an actual Spark cluster. We describe the design, implementation, and performance of Flint, along with the challenges associated with serverless analytics.

cs.DC↗

arXiv2018

Adaptive Pruning of Neural Language Models for Mobile Devices

Neural language models (NLMs) exist in an accuracy-efficiency tradeoff space where better perplexity typically comes at the cost of greater computation complexity. In a software keyboard application on mobile devices, this translates into higher power consumption and shorter battery life. This paper represents the first attempt, to our knowledge, in exploring accuracy-efficiency tradeoffs for NLMs. Building on quasi-recurrent neural networks (QRNNs), we apply pruning techniques to provide a "knob" to select different operating points. In addition, we propose a simple technique to recover some perplexity using a negligible amount of memory. Our empirical evaluations consider both perplexity as well as energy consumption on a Raspberry Pi, where we demonstrate which methods provide the best perplexity-power consumption operating point. At one operating point, one of the techniques is able to provide energy savings of 40% over the state of the art with only a 17% relative increase in perplexity.

cs.CL↗

arXiv2018

Deep Residual Learning for Small-Footprint Keyword Spotting

We explore the application of deep residual learning and dilated convolutions to the keyword spotting task, using the recently-released Google Speech Commands Dataset as our benchmark. Our best residual network (ResNet) implementation significantly outperforms Google's previous convolutional neural networks in terms of accuracy. By varying model depth and width, we can achieve compact models that also outperform previous small-footprint variants. To our knowledge, we are the first to examine these approaches for keyword spotting, and our results establish an open-source state-of-the-art reference to support the development of future speech-based interfaces.

cs.CL↗

arXiv2018

An Experimental Analysis of the Power Consumption of Convolutional Neural Networks for Keyword Spotting

Nearly all previous work on small-footprint keyword spotting with neural networks quantify model footprint in terms of the number of parameters and multiply operations for a feedforward inference pass. These values are, however, proxy measures since empirical performance in actual deployments is determined by many factors. In this paper, we study the power consumption of a family of convolutional neural networks for keyword spotting on a Raspberry Pi. We find that both proxies are good predictors of energy usage, although the number of multiplies is more predictive than the number of model parameters. We also confirm that models with the highest accuracies are, unsurprisingly, the most power hungry.

cs.OH↗

arXiv2018

Sapphire: Querying RDF Data Made Simple

RDF data in the linked open data (LOD) cloud is very valuable for many different applications. In order to unlock the full value of this data, users should be able to issue complex queries on the RDF datasets in the LOD cloud. SPARQL can express such complex queries, but constructing SPARQL queries can be a challenge to users since it requires knowing the structure and vocabulary of the datasets being queried. In this paper, we introduce Sapphire, a tool that helps users write syntactically and semantically correct SPARQL queries without prior knowledge of the queried datasets. Sapphire interactively helps the user while typing the query by providing auto-complete suggestions based on the queried data. After a query is issued, Sapphire provides suggestions on ways to change the query to better match the needs of the user. We evaluated Sapphire based on performance experiments and a user study and showed it to be superior to competing approaches.

cs.DB↗

arXiv2018

Strong Baselines for Simple Question Answering over Knowledge Graphs with and without Neural Networks

We examine the problem of question answering over knowledge graphs, focusing on simple questions that can be answered by the lookup of a single fact. Adopting a straightforward decomposition of the problem into entity detection, entity linking, relation prediction, and evidence combination, we explore simple yet strong baselines. On the popular SimpleQuestions dataset, we find that basic LSTMs and GRUs plus a few heuristics yield accuracies that approach the state of the art, and techniques that do not use neural networks also perform reasonably well. These results show that gains from sophisticated deep learning techniques proposed in the literature are quite modest and that some previous models exhibit unnecessary complexity.

cs.CL↗

arXiv2018

In-Browser Split-Execution Support for Interactive Analytics in the Cloud

The canonical analytics architecture today consists of a browser connected to a backend in the cloud. In all deployments that we are aware of, the browser is simply a dumb rendering endpoint. As an alternative, this paper explores split-execution architectures that push analytics capabilities into the browser. We show that, by taking advantage of typed arrays and asm.js, it is possible to build an analytical RDBMS in JavaScript that runs in a browser, achieving performance rivaling native databases. To support interactive data exploration, our Afterburner prototype automatically generates local materialized views from a backend database that are then shipped to the browser to facilitate subsequent interactions seamlessly and efficiently. We compare this architecture to several alternative deployments, experimentally demonstrating performance parity, while at the same time providing additional advantages in terms of administrative and operational simplicity.

cs.DB↗

arXiv2017

Honk: A PyTorch Reimplementation of Convolutional Neural Networks for Keyword Spotting

We describe Honk, an open-source PyTorch reimplementation of convolutional neural networks for keyword spotting that are included as examples in TensorFlow. These models are useful for recognizing "command triggers" in speech-based interfaces (e.g., "Hey Siri"), which serve as explicit cues for audio recordings of utterances that are sent to the cloud for full speech recognition. Evaluation on Google's recently released Speech Commands Dataset shows that our reimplementation is comparable in accuracy and provides a starting point for future work on the keyword spotting task.

cs.CL↗

arXiv2017

An Exploration of Approaches to Integrating Neural Reranking Models in Multi-Stage Ranking Architectures

We explore different approaches to integrating a simple convolutional neural network (CNN) with the Lucene search engine in a multi-stage ranking architecture. Our models are trained using the PyTorch deep learning toolkit, which is implemented in C/C++ with a Python frontend. One obvious integration strategy is to expose the neural network directly as a service. For this, we use Apache Thrift, a software framework for building scalable cross-language services. In exploring alternative architectures, we observe that once trained, the feedforward evaluation of neural networks is quite straightforward. Therefore, we can extract the parameters of a trained CNN from PyTorch and import the model into Java, taking advantage of the Java Deeplearning4J library for feedforward evaluation. This has the advantage that the entire end-to-end system can be implemented in Java. As a third approach, we can extract the neural network from PyTorch and "compile" it into a C++ program that exposes a Thrift service. We evaluate these alternatives in terms of performance (latency and throughput) as well as ease of integration. Experiments show that feedforward evaluation of the convolutional neural network is significantly slower in Java, while the performance of the compiled C++ network does not consistently beat the PyTorch implementation.

cs.IR↗

arXiv2017

Integrating Lexical and Temporal Signals in Neural Ranking Models for Searching Social Media Streams

Time is an important relevance signal when searching streams of social media posts. The distribution of document timestamps from the results of an initial query can be leveraged to infer the distribution of relevant documents, which can then be used to rerank the initial results. Previous experiments have shown that kernel density estimation is a simple yet effective implementation of this idea. This paper explores an alternative approach to mining temporal signals with recurrent neural networks. Our intuition is that neural networks provide a more expressive framework to capture the temporal coherence of neighboring documents in time. To our knowledge, we are the first to integrate lexical and temporal signals in an end-to-end neural network architecture, in which existing neural ranking models are used to generate query-document similarity vectors that feed into a bidirectional LSTM layer for temporal modeling. Our results are mixed: existing neural models for document ranking alone yield limited improvements over simple baselines, but the integration of lexical and temporal signals yield significant improvements over competitive temporal baselines.

cs.IR↗