arXiv subjects

Michael Collins

Publications and source records attributed to Michael Collins.

At least 37 records · Page 2Linked to original sources

arXiv2022

Measuring Attribution in Natural Language Generation Models

With recent improvements in natural language generation (NLG) models for various applications, it has become imperative to have the means to identify and evaluate whether NLG output is only sharing verifiable information about the external world. In this work, we present a new evaluation framework entitled Attributable to Identified Sources (AIS) for assessing the output of natural language generation models, when such output pertains to the external world. We first define AIS and introduce a two-stage annotation pipeline for allowing annotators to appropriately evaluate model output according to AIS guidelines. We empirically validate this approach on generation datasets spanning three tasks (two conversational QA datasets, a summarization dataset, and a table-to-text dataset) via human evaluation studies that suggest that AIS could serve as a common framework for measuring whether model-generated statements are supported by underlying sources. We release guidelines for the human evaluation studies.

cs.CL↗

arXiv2022

A Well-Composed Text is Half Done! Composition Sampling for Diverse Conditional Generation

We propose Composition Sampling, a simple but effective method to generate diverse outputs for conditional generation of higher quality compared to previous stochastic decoding strategies. It builds on recently proposed plan-based neural generation models (Narayan et al, 2021) that are trained to first create a composition of the output and then generate by conditioning on it and the input. Our approach avoids text degeneration by first sampling a composition in the form of an entity chain and then using beam search to generate the best possible text grounded to this entity chain. Experiments on summarization (CNN/DailyMail and XSum) and question generation (SQuAD), using existing and newly proposed automatic metrics together with human-based evaluation, demonstrate that Composition Sampling is currently the best available decoding strategy for generating diverse meaningful outputs.

cs.CL↗

arXiv2022

High-Fidelity Ion State Detection Using Trap-Integrated Avalanche Photodiodes

Integrated technologies greatly enhance the prospects for practical quantum information processing and sensing devices based on trapped ions. High-speed and high-fidelity ion state readout is critical for any such application. Integrated detectors offer significant advantages for system portability and can also greatly facilitate parallel operations if a separate detector can be incorporated at each ion-trapping location. Here we demonstrate ion quantum state detection at room temperature utilizing single-photon avalanche diodes (SPADs) integrated directly into the substrate of silicon ion trapping chips. We detect the state of a trapped $^{88}\text{Sr}^{+}$ ion via fluorescence collection with the SPAD, achieving $99.92(1)\%$ average fidelity in 450 $μ$s, opening the door to the application of integrated state detection to quantum computing and sensing utilizing arrays of trapped ions.

quant-ph↗

arXiv2021

Evaluating Explanations: How much do explanations from the teacher aid students?

While many methods purport to explain predictions by highlighting salient features, what aims these explanations serve and how they ought to be evaluated often go unstated. In this work, we introduce a framework to quantify the value of explanations via the accuracy gains that they confer on a student model trained to simulate a teacher model. Crucially, the explanations are available to the student during training, but are not available at test time. Compared to prior proposals, our approach is less easily gamed, enabling principled, automatic, model-agnostic evaluation of attributions. Using our framework, we compare numerous attribution methods for text classification and question answering, and observe quantitative differences that are consistent (to a moderate to high degree) across different student model architectures and learning strategies.

cs.CL↗

arXiv2021

NeurIPS 2020 EfficientQA Competition: Systems, Analyses and Lessons Learned

We review the EfficientQA competition from NeurIPS 2020. The competition focused on open-domain question answering (QA), where systems take natural language questions as input and return natural language answers. The aim of the competition was to build systems that can predict correct answers while also satisfying strict on-disk memory budgets. These memory budgets were designed to encourage contestants to explore the trade-off between storing retrieval corpora or the parameters of learned models. In this report, we describe the motivation and organization of the competition, review the best submissions, and analyze system predictions to inform a discussion of evaluation for open-domain QA.

cs.CL↗

arXiv2021

Partially Supervised Named Entity Recognition via the Expected Entity Ratio Loss

We study learning named entity recognizers in the presence of missing entity annotations. We approach this setting as tagging with latent variables and propose a novel loss, the Expected Entity Ratio, to learn models in the presence of systematically missing tags. We show that our approach is both theoretically sound and empirically useful. Experimentally, we find that it meets or exceeds performance of strong and state-of-the-art baselines across a variety of languages, annotation scenarios, and amounts of labeled data. In particular, we find that it significantly outperforms the previous state-of-the-art methods from Mayhew et al. (2019) and Li et al. (2021) by +12.7 and +2.3 F1 score in a challenging setting with only 1,000 biased annotations, averaged across 7 datasets. We also show that, when combined with our approach, a novel sparse annotation scheme outperforms exhaustive annotation for modest annotation budgets.

cs.CL↗

arXiv2021

Existence of parabolic minimizers to the total variation flow on metric measure spaces

We give an existence proof for variational solutions $u$ associated to the total variation flow. Here, the functions being considered are defined on a metric measure space $(\mathcal{X}, d, μ)$ satisfying a doubling condition and supporting a Poincaré inequality. For such parabolic minimizers that coincide with a time-independent Cauchy-Dirichlet datum $u_0$ on the parabolic boundary of a space-time-cylinder $Ω\times (0, T)$ with $Ω\subset \mathcal{X}$ an open set and $T > 0$, we prove existence in the weak parabolic function space $L^1_w(0, T; \mathrm{BV}(Ω))$. In this paper, we generalize results from a previous work by Bögelein, Duzaar and Marcellini by introducing a more abstract notion for $\mathrm{BV}$-valued parabolic function spaces. We argue completely on a variational level.

math.AP↗

arXiv2021

On planetary systems as ordered sequences

A planetary system consists of a host star and one or more planets, arranged into a particular configuration. Here, we consider what information belongs to the configuration, or ordering, of 4286 Kepler planets in their 3277 planetary systems. First, we train a neural network model to predict the radius and period of a planet based on the properties of its host star and the radii and period of its neighbors. The mean absolute error of the predictions of the trained model is a factor of 2.1 better than the MAE of the predictions of a naive model which draws randomly from dynamically allowable periods and radii. Second, we adapt a model used for unsupervised part-of-speech tagging in computational linguistics to investigate whether planets or planetary systems fall into natural categories with physically interpretable "grammatical rules." The model identifies two robust groups of planetary systems: (1) compact multi-planet systems and (2) systems around giant stars ($\log{g} \lesssim 4.0$), although the latter group is strongly sculpted by the selection bias of the transit method. These results reinforce the idea that planetary systems are not random sequences -- instead, as a population, they contain predictable patterns that can provide insight into the formation and evolution of planetary systems.

astro-ph.EP↗

arXiv2021

Sparse, Dense, and Attentional Representations for Text Retrieval

Dual encoders perform retrieval by encoding documents and queries into dense lowdimensional vectors, scoring each document by its inner product with the query. We investigate the capacity of this architecture relative to sparse bag-of-words models and attentional neural networks. Using both theoretical and empirical analysis, we establish connections between the encoding dimension, the margin between gold and lower-ranked documents, and the document length, suggesting limitations in the capacity of fixed-length encodings to support precise retrieval of long documents. Building on these insights, we propose a simple neural model that combines the efficiency of dual encoders with some of the expressiveness of more costly attentional architectures, and explore sparse-dense hybrids to capitalize on the precision of sparse retrieval. These models outperform strong alternatives in large-scale retrieval.

cs.CL↗

arXiv2021

Decontextualization: Making Sentences Stand-Alone

Models for question answering, dialogue agents, and summarization often interpret the meaning of a sentence in a rich context and use that meaning in a new context. Taking excerpts of text can be problematic, as key pieces may not be explicit in a local window. We isolate and define the problem of sentence decontextualization: taking a sentence together with its context and rewriting it to be interpretable out of context, while preserving its meaning. We describe an annotation procedure, collect data on the Wikipedia corpus, and use the data to train models to automatically decontextualize sentences. We present preliminary studies that show the value of sentence decontextualization in a user facing task, and as preprocessing for systems that perform document understanding. We argue that decontextualization is an important subtask in many downstream applications, and that the definitions and resources provided can benefit tasks that operate on sentences that occur in a richer context.

cs.CL↗

arXiv2020

QED: A Framework and Dataset for Explanations in Question Answering

A question answering system that in addition to providing an answer provides an explanation of the reasoning that leads to that answer has potential advantages in terms of debuggability, extensibility and trust. To this end, we propose QED, a linguistically informed, extensible framework for explanations in question answering. A QED explanation specifies the relationship between a question and answer according to formal semantic notions such as referential equality, sentencehood, and entailment. We describe and publicly release an expert-annotated dataset of QED explanations built upon a subset of the Google Natural Questions dataset, and report baseline models on two tasks -- post-hoc explanation generation given an answer, and joint question answering and explanation generation. In the joint setting, a promising result suggests that training on a relatively small amount of QED data can improve question answering. In addition to describing the formal, language-theoretic motivations for the QED approach, we describe a large user study showing that the presence of QED explanations significantly improves the ability of untrained raters to spot errors made by a strong neural QA baseline.

cs.CL↗

arXiv2020

TyDi QA: A Benchmark for Information-Seeking Question Answering in Typologically Diverse Languages

Confidently making progress on multilingual modeling requires challenging, trustworthy evaluations. We present TyDi QA---a question answering dataset covering 11 typologically diverse languages with 204K question-answer pairs. The languages of TyDi QA are diverse with regard to their typology---the set of linguistic features each language expresses---such that we expect models performing well on this set to generalize across a large number of the world's languages. We present a quantitative analysis of the data quality and example-level qualitative linguistic analyses of observed language phenomena that would not be found in English-only corpora. To provide a realistic information-seeking task and avoid priming effects, questions are written by people who want to know the answer, but don't know the answer yet, and the data is collected directly in each language without the use of translation.

cs.CL↗

arXiv2019

A BERT Baseline for the Natural Questions

This technical note describes a new baseline for the Natural Questions. Our model is based on BERT and reduces the gap between the model F1 scores reported in the original dataset paper and the human upper bound by 30% and 50% relative for the long and short answer tasks respectively. This baseline has been submitted to the official NQ leaderboard at ai.google.com/research/NaturalQuestions. Code, preprocessed data and pretrained model are available at https://github.com/google-research/language/tree/master/language/question_answering/bert_joint.

cs.CL↗

arXiv2019

Fusion of Detected Objects in Text for Visual Question Answering

To advance models of multimodal context, we introduce a simple yet powerful neural architecture for data that combines vision and natural language. The "Bounding Boxes in Text Transformer" (B2T2) also leverages referential information binding words to portions of the image in a single unified architecture. B2T2 is highly effective on the Visual Commonsense Reasoning benchmark (https://visualcommonsense.com), achieving a new state-of-the-art with a 25% relative reduction in error rate compared to published baselines and obtaining the best performance to date on the public leaderboard (as of May 22, 2019). A detailed ablation analysis shows that the early integration of the visual features into the text analysis is key to the effectiveness of the new architecture. A reference implementation of our models is provided (https://github.com/google-research/language/tree/master/language/question_answering/b2t2).

cs.CL↗

arXiv2019

The multiplicity distribution of Kepler's exoplanets

The true multiplicity distribution of transiting planet systems is obscured by strong observational biases, leading low-multiplicity systems to be overrepresented in the observed sample. Using the Kepler FGK planet hosts, we employ approximate Bayesian computation to infer the multiplicity distribution by comparing simulated catalogs to the observed one. After comparing a total of ten different multiplicity distributions, half of which were two-population models, to the observed data, we find that a single-population model following a Zipfian distribution is able to explain the Kepler data as well as any of the dichotomous models we test. Our work provides another example of a way to explain the observed Kepler multiplicities without invoking a dichotomous planet population. Using our preferred Zipfian model, we estimate that an additional $2393_{-717}^{+904}$ planets likely reside in the 1537 FGK Kepler systems studied in this work, which would increase the planet count by a factor of $2.22_{-0.36}^{+0.46}$. Of these hidden worlds, $663_{-151}^{+158}$ are expected to reside in ostensibly single-transiting-planet systems, meaning that an additional planet(s) is expected for approximately 1-in-2 such Kepler systems.

astro-ph.EP↗

arXiv2019

Synthetic QA Corpora Generation with Roundtrip Consistency

We introduce a novel method of generating synthetic question answering corpora by combining models of question generation and answer extraction, and by filtering the results to ensure roundtrip consistency. By pretraining on the resulting corpora we obtain significant improvements on SQuAD2 and NQ, establishing a new state-of-the-art on the latter. Our synthetic data generation models, for both question generation and answer extraction, can be fully reproduced by finetuning a publicly available BERT model on the extractive subsets of SQuAD2 and NQ. We also describe a more powerful variant that does full sequence-to-sequence pretraining for question generation, obtaining exact match and F1 at less than 0.1% and 0.4% from human performance on SQuAD2.

cs.CL↗

arXiv2019

BoolQ: Exploring the Surprising Difficulty of Natural Yes/No Questions

In this paper we study yes/no questions that are naturally occurring --- meaning that they are generated in unprompted and unconstrained settings. We build a reading comprehension dataset, BoolQ, of such questions, and show that they are unexpectedly challenging. They often query for complex, non-factoid information, and require difficult entailment-like inference to solve. We also explore the effectiveness of a range of transfer learning baselines. We find that transferring from entailment data is more effective than transferring from paraphrase or extractive QA data, and that it, surprisingly, continues to be very beneficial even when starting from massive pre-trained language models such as BERT. Our best method trains BERT on MultiNLI and then re-trains it on our train set. It achieves 80.4% accuracy compared to 90% accuracy of human annotators (and 62% majority-baseline), leaving a significant gap for future work.

cs.CL↗

arXiv2019

Low-Resource Syntactic Transfer with Unsupervised Source Reordering

We describe a cross-lingual transfer method for dependency parsing that takes into account the problem of word order differences between source and target languages. Our model only relies on the Bible, a considerably smaller parallel data than the commonly used parallel data in transfer methods. We use the concatenation of projected trees from the Bible corpus, and the gold-standard treebanks in multiple source languages along with cross-lingual word representations. We demonstrate that reordering the source treebanks before training on them for a target language improves the accuracy of languages outside the European language family. Our experiments on 68 treebanks (38 languages) in the Universal Dependencies corpus achieve a high accuracy for all languages. Among them, our experiments on 16 treebanks of 12 non-European languages achieve an average UAS absolute improvement of 3.3% over a state-of-the-art method.

cs.CL↗