We submitted two systems to the SemEval-2016 Task 12: Clinical TempEval challenge, participating in Phase 1, where we identified text spans of time and event expressions in clinical notes and Phase 2, where we predicted a relation between an event and its parent document creation time. For temporal entity extraction, we find that a joint inference-based approach using structured prediction outperforms a vanilla recurrent neural network that incorporates word embeddings trained on a variety of large clinical document sets. For document creation time relations, we find that a combination of date canonicalization and distant supervision rules for predicting relations on both events and time expressions improves classification, though gains are limited, likely due to the small scale of training data.
Margin-based structured prediction commonly uses a maximum loss over all possible structured outputs \cite{Altun03,Collins04b,Taskar03}. In natural language processing, recent work \cite{Zhang14,Zhang15} has proposed the use of the maximum loss over random structured outputs sampled independently from some proposal distribution. This method is linear-time in the number of random structured outputs and trivially parallelizable. We study this family of loss functions in the PAC-Bayes framework under Gaussian perturbations \cite{McAllester07}. Under some technical conditions and up to statistical accuracy, we show that this family of loss functions produces a tighter upper bound of the Gibbs decoder distortion than commonly used methods. Thus, using the maximum loss over random structured outputs is a principled way of learning the parameter of structured prediction models. Besides explaining the experimental success of \cite{Zhang14,Zhang15}, our theoretical results show that more general techniques are possible.
We study the problem of structured prediction under test-time budget constraints. We propose a novel approach applicable to a wide range of structured prediction problems in computer vision and natural language processing. Our approach seeks to adaptively generate computationally costly features during test-time in order to reduce the computational cost of prediction while maintaining prediction performance. We show that training the adaptive feature generation system can be reduced to a series of structured learning problems, resulting in efficient training using existing structured learning algorithms. This framework provides theoretical justification for several existing heuristic approaches found in literature. We evaluate our proposed adaptive system on two structured prediction tasks, optical character recognition (OCR) and dependency parsing and show strong performance in reduction of the feature costs without degrading accuracy.
Existing methods for pixel-wise labelling tasks generally disregard the underlying structure of labellings, often leading to predictions that are visually implausible. While incorporating structure into the model should improve prediction quality, doing so is challenging - manually specifying the form of structural constraints may be impractical and inference often becomes intractable even if structural constraints are given. We sidestep this problem by reducing structured prediction to a sequence of unconstrained prediction problems and demonstrate that this approach is capable of automatically discovering priors on shape, contiguity of region predictions and smoothness of region contours from data without any a priori specification. On the instance segmentation task, this method outperforms the state-of-the-art, achieving a mean $\mathrm{AP}^{r}$ of 63.6% at 50% overlap and 43.3% at 70% overlap.
Domain-general semantic parsing is a long-standing goal in natural language processing, where the semantic parser is capable of robustly parsing sentences from domains outside of which it was trained. Current approaches largely rely on additional supervision from new domains in order to generalize to those domains. We present a generative model of natural language utterances and logical forms and demonstrate its application to semantic parsing. Our approach relies on domain-independent supervision to generalize to new domains. We derive and implement efficient algorithms for training, parsing, and sentence generation. The work relies on a novel application of hierarchical Dirichlet processes (HDPs) for structured prediction, which we also present in this manuscript. This manuscript is an excerpt of chapter 4 from the Ph.D. thesis of Saparov (2022), where the model plays a central role in a larger natural language understanding system. This manuscript provides a new simplified and more complete presentation of the work first introduced in Saparov, Saraswat, and Mitchell (2017). The description and proofs of correctness of the training algorithm, parsing algorithm, and sentence generation algorithm are much simplified in this new presentation. We also describe the novel application of hierarchical Dirichlet processes for structured prediction. In addition, we extend the earlier work with a new model of word morphology, which utilizes the comprehensive morphological data from Wiktionary.
We introduce structured prediction energy networks (SPENs), a flexible framework for structured prediction. A deep architecture is used to define an energy function of candidate labels, and then predictions are produced by using back-propagation to iteratively optimize the energy with respect to the labels. This deep architecture captures dependencies between labels that would lead to intractable graphical models, and performs structure learning by automatically learning discriminative features of the structured output. One natural application of our technique is multi-label classification, which traditionally has required strict prior assumptions about the interactions between labels to ensure tractable learning and prediction. We are able to apply SPENs to multi-label problems with substantially larger label sets than previous applications of structured prediction, while modeling high-order interactions using minimal structural assumptions. Overall, deep learning provides remarkable tools for learning features of the inputs to a prediction problem, and this work extends these techniques to learning features of structured outputs. Our experiments provide impressive performance on a variety of benchmark multi-label classification tasks, demonstrate that our technique can be used to provide interpretable structure learning, and illuminate fundamental trade-offs between feed-forward and iterative structured prediction.
We propose a novel supervised learning technique for summarizing videos by automatically selecting keyframes or key subshots. Casting the problem as a structured prediction problem on sequential data, our main idea is to use Long Short-Term Memory (LSTM), a special type of recurrent neural networks to model the variable-range dependencies entailed in the task of video summarization. Our learning models attain the state-of-the-art results on two benchmark video datasets. Detailed analysis justifies the design of the models. In particular, we show that it is crucial to take into consideration the sequential structures in videos and model them. Besides advances in modeling techniques, we introduce techniques to address the need of a large number of annotated data for training complex learning models. There, our main idea is to exploit the existence of auxiliary annotated video datasets, albeit heterogeneous in visual styles and contents. Specifically, we show domain adaptation techniques can improve summarization by reducing the discrepancies in statistical properties across those datasets.
Sequence labeling is a widely used method for named entity recognition and information extraction from unstructured natural language data. In clinical domain one major application of sequence labeling involves extraction of medical entities such as medication, indication, and side-effects from Electronic Health Record narratives. Sequence labeling in this domain, presents its own set of challenges and objectives. In this work we experimented with various CRF based structured learning models with Recurrent Neural Networks. We extend the previously studied LSTM-CRF models with explicit modeling of pairwise potentials. We also propose an approximate version of skip-chain CRF inference with RNN potentials. We use these methodologies for structured prediction in order to improve the exact phrase detection of various medical entities.
Discriminative segmental models, such as segmental conditional random fields (SCRFs) and segmental structured support vector machines (SSVMs), have had success in speech recognition via both lattice rescoring and first-pass decoding. However, such models suffer from slow decoding, hampering the use of computationally expensive features, such as segment neural networks or other high-order features. A typical solution is to use approximate decoding, either by beam pruning in a single pass or by beam pruning to generate a lattice followed by a second pass. In this work, we study discriminative segmental models trained with a hinge loss (i.e., segmental structured SVMs). We show that beam search is not suitable for learning rescoring models in this approach, though it gives good approximate decoding performance when the model is already well-trained. Instead, we consider an approach inspired by structured prediction cascades, which use max-marginal pruning to generate lattices. We obtain a high-accuracy phonetic recognition system with several expensive feature types: a segment neural network, a second-order language model, and second-order phone boundary features.
In structured prediction problems where we have indirect supervision of the output, maximum marginal likelihood faces two computational obstacles: non-convexity of the objective and intractability of even a single gradient computation. In this paper, we bypass both obstacles for a class of what we call linear indirectly-supervised problems. Our approach is simple: we solve a linear system to estimate sufficient statistics of the model, which we then use to estimate parameters via convex optimization. We analyze the statistical properties of our approach and show empirically that it is effective in two settings: learning with local privacy constraints and learning from low-cost count-based annotations.
We propose a novel semi-supervised structured output prediction method based on local linear regression in this paper. The existing semi-supervise structured output prediction methods learn a global predictor for all the data points in a data set, which ignores the differences of local distributions of the data set, and the effects to the structured output prediction. To solve this problem, we propose to learn the missing structured outputs and local predictors for neighborhoods of different data points jointly. Using the local linear regression strategy, in the neighborhood of each data point, we propose to learn a local linear predictor by minimizing both the complexity of the predictor and the upper bound of the structured prediction loss. The minimization problem is solved by sub-gradient descent algorithms. We conduct experiments over two benchmark data sets, and the results show the advantages of the proposed method.
Recurrent neural networks have recently been used for learning to describe images using natural language. However, it has been observed that these models generalize poorly to scenes that were not observed during training, possibly depending too strongly on the statistics of the text in the training data. Here we propose to describe images using short structured representations, aiming to capture the crux of a description. These structured representations allow us to tease-out and evaluate separately two types of generalization: standard generalization to new images with similar scenes, and generalization to new combinations of known entities. We compare two learning approaches on the MS-COCO dataset: a state-of-the-art recurrent network based on an LSTM (Show, Attend and Tell), and a simple structured prediction model on top of a deep network. We find that the structured model generalizes to new compositions substantially better than the LSTM, ~7 times the accuracy of predicting structured representations. By providing a concrete method to quantify generalization for unseen combinations, we argue that structured representations and compositional splits are a useful benchmark for image captioning, and advocate compositional models that capture linguistic and visual structure.
We develop an automated variational inference method for Bayesian structured prediction problems with Gaussian process (GP) priors and linear-chain likelihoods. Our approach does not need to know the details of the structured likelihood model and can scale up to a large number of observations. Furthermore, we show that the required expected likelihood term and its gradients in the variational objective (ELBO) can be estimated efficiently by using expectations over very low-dimensional Gaussian distributions. Optimization of the ELBO is fully parallelizable over sequences and amenable to stochastic optimization, which we use along with control variate techniques and state-of-the-art incremental optimization to make our framework useful in practice. Results on a set of natural language processing tasks show that our method can be as good as (and sometimes better than) hard-coded approaches including SVM-struct and CRFs, and overcomes the scalability limitations of previous inference algorithms based on sampling. Overall, this is a fundamental step to developing automated inference methods for Bayesian structured prediction.
We present the Structured Weighted Violations Perceptron (SWVP) algorithm, a new structured prediction algorithm that generalizes the Collins Structured Perceptron (CSP). Unlike CSP, the update rule of SWVP explicitly exploits the internal structure of the predicted labels. We prove the convergence of SWVP for linearly separable training sets, provide mistake and generalization bounds, and show that in the general case these bounds are tighter than those of the CSP special case. In synthetic data experiments with data drawn from an HMM, various variants of SWVP substantially outperform its CSP special case. SWVP also provides encouraging initial dependency parsing results.
Entity linking is the task of identifying mentions of entities in text, and linking them to entries in a knowledge base. This task is especially difficult in microblogs, as there is little additional text to provide disambiguating context; rather, authors rely on an implicit common ground of shared knowledge with their readers. In this paper, we attempt to capture some of this implicit context by exploiting the social network structure in microblogs. We build on the theory of homophily, which implies that socially linked individuals share interests, and are therefore likely to mention the same sorts of entities. We implement this idea by encoding authors, mentions, and entities in a continuous vector space, which is constructed so that socially-connected authors have similar vector representations. These vectors are incorporated into a neural structured prediction model, which captures structural constraints that are inherent in the entity linking task. Together, these design decisions yield F1 improvements of 1%-5% on benchmark datasets, as compared to the previous state-of-the-art.
We describe an end-to-end framework for learning parameters of min-cost flow multi-target tracking problem with quadratic trajectory interactions including suppression of overlapping tracks and contextual cues about cooccurrence of different objects. Our approach utilizes structured prediction with a tracking-specific loss function to learn the complete set of model parameters. In this learning framework, we evaluate two different approaches to finding an optimal set of tracks under a quadratic model objective, one based on an LP relaxation and the other based on novel greedy variants of dynamic programming that handle pairwise interactions. We find the greedy algorithms achieve almost equivalent accuracy to the LP relaxation while being up to 10x faster than a commercial LP solver. We evaluate trained models on three challenging benchmarks. Surprisingly, we find that with proper parameter learning, our simple data association model without explicit appearance/motion reasoning is able to achieve comparable or better accuracy than many state-of-the-art methods that use far more complex motion features or appearance affinity metric learning.
Conditional random fields (CRFs) provide a powerful tool for structured prediction, but cast significant challenges in both the learning and inference steps. Approximation techniques are widely used in both steps, which should be considered jointly to guarantee good performance (a.k.a. "inferning"). Perturb-and-MAP models provide a promising alternative to CRFs, but require global combinatorial optimization and hence they are usable only on specific models. In this work, we present a new Local Perturb-and-MAP (locPMAP) framework that replaces the global optimization with a local optimization by exploiting our observed connection between locPMAP and the pseudolikelihood of the original CRF model. We test our approach on three different vision tasks and show that our method achieves consistently improved performance over other approximate inference techniques optimized to a pseudolikelihood objective. Additionally, we demonstrate that we can integrate our method in the fully convolutional network framework to increase our model's complexity. Finally, our observed connection between locPMAP and the pseudolikelihood leads to a novel perspective for understanding and using pseudolikelihood.
A key barrier to making phonetic studies scalable and replicable is the need to rely on subjective, manual annotation. To help meet this challenge, a machine learning algorithm was developed for automatic measurement of a widely used phonetic measure: vowel duration. Manually-annotated data were used to train a model that takes as input an arbitrary length segment of the acoustic signal containing a single vowel that is preceded and followed by consonants and outputs the duration of the vowel. The model is based on the structured prediction framework. The input signal and a hypothesized set of a vowel's onset and offset are mapped to an abstract vector space by a set of acoustic feature functions. The learning algorithm is trained in this space to minimize the difference in expectations between predicted and manually-measured vowel durations. The trained model can then automatically estimate vowel durations without phonetic or orthographic transcription. Results comparing the model to three sets of manually annotated data suggest it out-performed the current gold standard for duration measurement, an HMM-based forced aligner (which requires orthographic or phonetic transcription as an input).
We describe and analyze a simple and effective algorithm for sequence segmentation applied to speech processing tasks. We propose a neural architecture that is composed of two modules trained jointly: a recurrent neural network (RNN) module and a structured prediction model. The RNN outputs are considered as feature functions to the structured model. The overall model is trained with a structured loss function which can be designed to the given segmentation task. We demonstrate the effectiveness of our method by applying it to two simple tasks commonly used in phonetic studies: word segmentation and voice onset time segmentation. Results sug- gest the proposed model is superior to previous methods, ob- taining state-of-the-art results on the tested datasets.
Stochastic structured prediction under bandit feedback follows a learning protocol where on each of a sequence of iterations, the learner receives an input, predicts an output structure, and receives partial feedback in form of a task loss evaluation of the predicted structure. We present applications of this learning scenario to convex and non-convex objectives for structured prediction and analyze them as stochastic first-order methods. We present an experimental evaluation on problems of natural language processing over exponential output spaces, and compare convergence speed across different objectives under the practical criterion of optimal task performance on development data and the optimization-theoretic criterion of minimal squared gradient norm. Best results under both criteria are obtained for a non-convex objective for pairwise preference learning under bandit feedback.