arXiv subjects

Alex Wang

Publications and source records attributed to Alex Wang.

35 records · Page 2Linked to original sources

arXiv2022

Benchmarking Active Learning Strategies for Materials Optimization and Discovery

Autonomous physical science is revolutionizing materials science. In these systems, machine learning controls experiment design, execution, and analysis in a closed loop. Active learning, the machine learning field of optimal experiment design, selects each subsequent experiment to maximize knowledge toward the user goal. Autonomous system performance can be further improved with implementation of scientific machine learning, also known as inductive bias-engineered artificial intelligence, which folds prior knowledge of physical laws (e.g., Gibbs phase rule) into the algorithm. As the number, diversity, and uses for active learning strategies grow, there is an associated growing necessity for real-world reference datasets to benchmark strategies. We present a reference dataset and demonstrate its use to benchmark active learning strategies in the form of various acquisition functions. Active learning strategies are used to rapidly identify materials with optimal physical properties within a ternary materials system. The data is from an actual Fe-Co-Ni thin-film library and includes previously acquired experimental data for materials compositions, X-ray diffraction patterns, and two functional properties of magnetic coercivity and the Kerr rotation. Popular active learning methods along with a recent scientific active learning method are benchmarked for their materials optimization performance. We discuss the relationship between algorithm performance, materials search space complexity, and the incorporation of prior knowledge.

cond-mat.mtrl-sci↗

arXiv2021

Application of Quantum Machine Learning using the Quantum Variational Classifier Method to High Energy Physics Analysis at the LHC on IBM Quantum Computer Simulator and Hardware with 10 qubits

One of the major objectives of the experimental programs at the LHC is the discovery of new physics. This requires the identification of rare signals in immense backgrounds. Using machine learning algorithms greatly enhances our ability to achieve this objective. With the progress of quantum technologies, quantum machine learning could become a powerful tool for data analysis in high energy physics. In this study, using IBM gate-model quantum computing systems, we employ the quantum variational classifier method in two recent LHC flagship physics analyses: $t\bar{t}H$ (Higgs boson production in association with a top quark pair) and $H\rightarrowμ^{+}μ^{-}$ (Higgs boson decays to two muons, probing the Higgs boson couplings to second-generation fermions). We have obtained early results with 10 qubits on the IBM quantum simulator and the IBM quantum hardware. With small training samples of 100 events on the quantum simulator, the quantum variational classifier method performs similarly to classical algorithms such as SVM (support vector machine) and BDT (boosted decision tree), which are often employed in LHC physics analyses. On the quantum hardware, the quantum variational classifier method has shown promising discrimination power, comparable to that on the quantum simulator. This study demonstrates that quantum machine learning has the ability to differentiate between signal and background in realistic physics datasets. We foresee the usage of quantum machine learning in future high-luminosity LHC physics analyses, including measurements of the Higgs boson self-couplings and searches for dark matter.

quant-ph↗

arXiv2021

QuestEval: Summarization Asks for Fact-based Evaluation

Summarization evaluation remains an open research problem: current metrics such as ROUGE are known to be limited and to correlate poorly with human judgments. To alleviate this issue, recent work has proposed evaluation metrics which rely on question answering models to assess whether a summary contains all the relevant information in its source document. Though promising, the proposed approaches have so far failed to correlate better than ROUGE with human judgments. In this paper, we extend previous approaches and propose a unified framework, named QuestEval. In contrast to established metrics such as ROUGE or BERTScore, QuestEval does not require any ground-truth reference. Nonetheless, QuestEval substantially improves the correlation with human judgments over four evaluation dimensions (consistency, coherence, fluency, and relevance), as shown in the extensive experiments we report.

cs.CL↗

arXiv2020

jiant: A Software Toolkit for Research on General-Purpose Text Understanding Models

We introduce jiant, an open source toolkit for conducting multitask and transfer learning experiments on English NLU tasks. jiant enables modular and configuration-driven experimentation with state-of-the-art models and implements a broad set of tasks for probing, transfer learning, and multitask training experiments. jiant implements over 50 NLU tasks, including all GLUE and SuperGLUE benchmark tasks. We demonstrate that jiant reproduces published performance on a variety of tasks and models, including BERT and RoBERTa. jiant is available at https://jiant.info.

cs.CL↗

arXiv2020

Asking and Answering Questions to Evaluate the Factual Consistency of Summaries

Practical applications of abstractive summarization models are limited by frequent factual inconsistencies with respect to their input. Existing automatic evaluation metrics for summarization are largely insensitive to such errors. We propose an automatic evaluation protocol called QAGS (pronounced "kags") that is designed to identify factual inconsistencies in a generated summary. QAGS is based on the intuition that if we ask questions about a summary and its source, we will receive similar answers if the summary is factually consistent with the source. To evaluate QAGS, we collect human judgments of factual consistency on model-generated summaries for the CNN/DailyMail (Hermann et al., 2015) and XSUM (Narayan et al., 2018) summarization datasets. QAGS has substantially higher correlations with these judgments than other automatic evaluation metrics. Also, QAGS offers a natural form of interpretability: The answers and questions generated while computing QAGS indicate which tokens of a summary are inconsistent and why. We believe QAGS is a promising tool in automatically generating usable and factually consistent text.

cs.CL↗

arXiv2020

SuperGLUE: A Stickier Benchmark for General-Purpose Language Understanding Systems

In the last year, new models and methods for pretraining and transfer learning have driven striking performance improvements across a range of language understanding tasks. The GLUE benchmark, introduced a little over one year ago, offers a single-number metric that summarizes progress on a diverse set of such tasks, but performance on the benchmark has recently surpassed the level of non-expert humans, suggesting limited headroom for further research. In this paper we present SuperGLUE, a new benchmark styled after GLUE with a new set of more difficult language understanding tasks, a software toolkit, and a public leaderboard. SuperGLUE is available at super.gluebenchmark.com.

cs.CL↗

arXiv2020

A Generalized Framework of Sequence Generation with Application to Undirected Sequence Models

Undirected neural sequence models such as BERT (Devlin et al., 2019) have received renewed interest due to their success on discriminative natural language understanding tasks such as question-answering and natural language inference. The problem of generating sequences directly from these models has received relatively little attention, in part because generating from undirected models departs significantly from conventional monotonic generation in directed sequence models. We investigate this problem by proposing a generalized model of sequence generation that unifies decoding in directed and undirected models. The proposed framework models the process of generation rather than the resulting sequence, and under this framework, we derive various neural sequence models as special cases, such as autoregressive, semi-autoregressive, and refinement-based non-autoregressive models. This unification enables us to adapt decoding algorithms originally developed for directed sequence models to undirected sequence models. We demonstrate this by evaluating various handcrafted and learned decoding strategies on a BERT-like machine translation model (Lample & Conneau, 2019). The proposed approach achieves constant-time translation results on par with linear-time translation results from the same undirected sequence model, while both are competitive with the state-of-the-art on WMT'14 English-German translation.

cs.LG↗

arXiv2019

Deep Learning for the Digital Pathologic Diagnosis of Cholangiocarcinoma and Hepatocellular Carcinoma: Evaluating the Impact of a Web-based Diagnostic Assistant

While artificial intelligence (AI) algorithms continue to rival human performance on a variety of clinical tasks, the question of how best to incorporate these algorithms into clinical workflows remains relatively unexplored. We investigated how AI can affect pathologist performance on the task of differentiating between two subtypes of primary liver cancer, hepatocellular carcinoma (HCC) and cholangiocarcinoma (CC). We developed an AI diagnostic assistant using a deep learning model and evaluated its effect on the diagnostic performance of eleven pathologists with varying levels of expertise. Our deep learning model achieved an accuracy of 0.885 on an internal validation set of 26 slides and an accuracy of 0.842 on an independent test set of 80 slides. Despite having high accuracy on a hold out test set, the diagnostic assistant did not significantly improve performance across pathologists (p-value: 0.184, OR: 1.287 (95% CI 0.886, 1.871)). Model correctness was observed to significantly bias the pathologist decisions. When the model was correct, assistance significantly improved accuracy across all pathologist experience levels and for all case difficulty levels (p-value: < 0.001, OR: 4.289 (95% CI 2.360, 7.794)). When the model was incorrect, assistance significantly decreased accuracy across all 11 pathologists and for all case difficulty levels (p-value < 0.001, OR: 0.253 (95% CI 0.126, 0.507)). Our results highlight the challenges of translating AI models to the clinical setting, especially for difficult subspecialty tasks such as tumor classification. In particular, they suggest that incorrect model predictions could strongly bias an expert's diagnosis, an important factor to consider when designing medical AI-assistance systems.

eess.IV↗

arXiv2019

Probing What Different NLP Tasks Teach Machines about Function Word Comprehension

We introduce a set of nine challenge tasks that test for the understanding of function words. These tasks are created by structurally mutating sentences from existing datasets to target the comprehension of specific types of function words (e.g., prepositions, wh-words). Using these probing tasks, we explore the effects of various pretraining objectives for sentence encoders (e.g., language modeling, CCG supertagging and natural language inference (NLI)) on the learned representations. Our results show that pretraining on language modeling performs the best on average across our probing tasks, supporting its widespread use for pretraining state-of-the-art NLP models, and CCG supertagging and NLI pretraining perform comparably. Overall, no pretraining objective dominates across the board, and our function word probing tasks highlight several intuitive differences between pretraining objectives, e.g., that NLI helps the comprehension of negation.

cs.CL↗

arXiv2019

Can You Tell Me How to Get Past Sesame Street? Sentence-Level Pretraining Beyond Language Modeling

Natural language understanding has recently seen a surge of progress with the use of sentence encoders like ELMo (Peters et al., 2018a) and BERT (Devlin et al., 2019) which are pretrained on variants of language modeling. We conduct the first large-scale systematic study of candidate pretraining tasks, comparing 19 different tasks both as alternatives and complements to language modeling. Our primary results support the use language modeling, especially when combined with pretraining on additional labeled-data tasks. However, our results are mixed across pretraining tasks and show some concerning trends: In ELMo's pretrain-then-freeze paradigm, random baselines are worryingly strong and results vary strikingly across target tasks. In addition, fine-tuning BERT on an intermediate task often negatively impacts downstream transfer. In a more positive trend, we see modest gains from multitask training, suggesting the development of more sophisticated multitask and transfer learning techniques as an avenue for further research.

cs.CL↗

arXiv2019

What do you learn from context? Probing for sentence structure in contextualized word representations

Contextualized representation models such as ELMo (Peters et al., 2018a) and BERT (Devlin et al., 2018) have recently achieved state-of-the-art results on a diverse array of downstream NLP tasks. Building on recent token-level probing work, we introduce a novel edge probing task design and construct a broad suite of sub-sentence tasks derived from the traditional structured NLP pipeline. We probe word-level contextual representations from four recent models and investigate how they encode sentence structure across a range of syntactic, semantic, local, and long-range phenomena. We find that existing models trained on language modeling and translation produce strong representations for syntactic phenomena, but only offer comparably small improvements on semantic tasks over a non-contextual baseline.

cs.CL↗

arXiv2019

BERT has a Mouth, and It Must Speak: BERT as a Markov Random Field Language Model

We show that BERT (Devlin et al., 2018) is a Markov random field language model. This formulation gives way to a natural procedure to sample sentences from BERT. We generate from BERT and find that it can produce high-quality, fluent generations. Compared to the generations of a traditional left-to-right language model, BERT generates sentences that are more diverse but of slightly worse quality.

cs.CL↗

arXiv2019

On Measuring Social Biases in Sentence Encoders

The Word Embedding Association Test shows that GloVe and word2vec word embeddings exhibit human-like implicit biases based on gender, race, and other social constructs (Caliskan et al., 2017). Meanwhile, research on learning reusable text representations has begun to explore sentence-level texts, with some sentence encoders seeing enthusiastic adoption. Accordingly, we extend the Word Embedding Association Test to measure bias in sentence encoders. We then test several sentence encoders, including state-of-the-art methods such as ELMo and BERT, for the social biases studied in prior work and two important biases that are difficult or impossible to test at the word level. We observe mixed results including suspicious patterns of sensitivity that suggest the test's assumptions may not hold in general. We conclude by proposing directions for future work on measuring bias in sentence encoders.

cs.CL↗

arXiv2019

GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding

For natural language understanding (NLU) technology to be maximally useful, both practically and as a scientific object of study, it must be general: it must be able to process language in a way that is not exclusively tailored to any one specific task or dataset. In pursuit of this objective, we introduce the General Language Understanding Evaluation benchmark (GLUE), a tool for evaluating and analyzing the performance of models across a diverse range of existing NLU tasks. GLUE is model-agnostic, but it incentivizes sharing knowledge across tasks because certain tasks have very limited training data. We further provide a hand-crafted diagnostic test suite that enables detailed linguistic analysis of NLU models. We evaluate baselines based on current methods for multi-task and transfer learning and find that they do not immediately give substantial improvements over the aggregate performance of training a separate model per task, indicating room for improvement in developing general and robust NLU systems.

cs.CL↗

arXiv2017

Clustering Stable Instances of Euclidean k-means

The Euclidean k-means problem is arguably the most widely-studied clustering problem in machine learning. While the k-means objective is NP-hard in the worst-case, practitioners have enjoyed remarkable success in applying heuristics like Lloyd's algorithm for this problem. To address this disconnect, we study the following question: what properties of real-world instances will enable us to design efficient algorithms and prove guarantees for finding the optimal clustering? We consider a natural notion called additive perturbation stability that we believe captures many practical instances. Stable instances have unique optimal k-means solutions that do not change even when each point is perturbed a little (in Euclidean distance). This captures the property that the k-means optimal solution should be tolerant to measurement errors and uncertainty in the points. We design efficient algorithms that provably recover the optimal clustering for instances that are additive perturbation stable. When the instance has some additional separation, we show an efficient algorithm with provable guarantees that is also robust to outliers. We complement these results by studying the amount of stability in real datasets and demonstrating that our algorithm performs well on these benchmark datasets.

cs.LG↗

arXiv2011

Intent Inference and Syntactic Tracking with GMTI Measurements

In conventional target tracking systems, human operators use the estimated target tracks to make higher level inference of the target behaviour/intent. This paper develops syntactic filtering algorithms that assist human operators by extracting spatial patterns from target tracks to identify suspicious/anomalous spatial trajectories. The targets' spatial trajectories are modeled by a stochastic context free grammar (SCFG) and a switched mode state space model. Bayesian filtering algorithms for stochastic context free grammars are presented for extracting the syntactic structure and illustrated for a ground moving target indicator (GMTI) radar example. The performance of the algorithms is tested with the experimental data collected using DRDC Ottawa's X-band Wideband Experimental Airborne Radar (XWEAR).

stat.ME↗

arXiv2005

Output Feedback Pole Assignment for Transfer Functions with Symmetries

This paper studies the problem of pole assignment for symmetric and Hamiltonian transfer functions. A necessary and sufficient condition for pole assignment by complex symmetric output feedback transformations is given. Moreover, in the case where the McMillan degree coincides with the number of parameters appearing in the symmetric feedback transformations, we derive an explicit combinatorial formula for the number of pole assigning symmetric feedback gains. The proof uses intersection theory in projective space as well as a formula for the degree of the complex Lagrangian Grassmann manifold.

math.OC↗