Item response theory (IRT) models for categorical response data are widely used in the analysis of educational data, computerized adaptive testing, and psychological surveys. However, most IRT models rely on both the assumption that categories are strictly ordered and the assumption that this ordering is known a priori. These assumptions are impractical in many real-world scenarios, such as multiple-choice exams where the levels of incorrectness for the distractor categories are often unknown. While a number of results exist on IRT models for unordered categorical data, they tend to have restrictive modeling assumptions that lead to poor data fitting performance in practice. Furthermore, existing unordered categorical models have parameters that are difficult to interpret. In this work, we propose a novel methodology for unordered categorical IRT that we call SPRITE (short for stochastic polytomous response item model) that: (i) analyzes both ordered and unordered categories, (ii) offers interpretable outputs, and (iii) provides improved data fitting compared to existing models. We compare SPRITE to existing item response models and demonstrate its efficacy on both synthetic and real-world educational datasets.

Coherent reasoning under uncertainty can be represented in a very general manner by coherent sets of desirable gambles. In a context that does not allow for indecision, this leads to an approach that is mathematically equivalent to working with coherent conditional probabilities. If we do allow for indecision, this leads to a more general foundation for coherent (imprecise-)probabilistic inference. In this framework, and for a given finite category set, coherent predictive inference under exchangeability can be represented using Bernstein coherent cones of multivariate polynomials on the simplex generated by this category set. This is a powerful generalisation of de Finetti's Representation Theorem allowing for both imprecision and indecision. We define an inference system as a map that associates a Bernstein coherent cone of polynomials with every finite category set. Many inference principles encountered in the literature can then be interpreted, and represented mathematically, as restrictions on such maps. We discuss, as particular examples, two important inference principles: representation insensitivitya strengthened version of Walley's representation invarianceand specificity. We show that there is an infinity of inference systems that satisfy these two principles, amongst which we discuss in particular the skeptically cautious inference system, the inference systems corresponding to (a modified version of) Walley and Bernard's Imprecise Dirichlet Multinomial Models (IDMM), the skeptical IDMM inference systems, and the Haldane inference system. We also prove that the latter produces the same posterior inferences as would be obtained using Haldane's improper prior, implying that there is an infinity of proper priors that produce the same coherent posterior inferences as Haldane's improper one. Finally, we impose an additional inference principle that allows us to characterise uniquely the immediate predictions for the IDMM inference systems.

While topic models can discover patterns of word usage in large corpora, it is difficult to meld this unsupervised structure with noisy, human-provided labels, especially when the label space is large. In this paper, we present a model-Label to Hierarchy (L2H)-that can induce a hierarchy of user-generated labels and the topics associated with those labels from a set of multi-labeled documents. The model is robust enough to account for missing labels from untrained, disparate annotators and provide an interpretable summary of an otherwise unwieldy label set. We show empirically the effectiveness of L2H in predicting held-out words and labels for unseen documents.

Distributed machine learning has typically been approached from a data parallel perspective, where big data are partitioned to multiple workers and an algorithm is executed concurrently over different data subsets under various synchronization schemes to ensure speed-up and/or correctness. A sibling problem that has received relatively less attention is how to ensure efficient and correct model parallel execution of ML algorithms, where parameters of an ML program are partitioned to different workers and undergone concurrent iterative updates. We argue that model and data parallelisms impose rather different challenges for system design, algorithmic adjustment, and theoretical analysis. In this paper, we develop a system for model-parallelism, STRADS, that provides a programming abstraction for scheduling parameter updates by discovering and leveraging changing structural properties of ML programs. STRADS enables a flexible tradeoff between scheduling efficiency and fidelity to intrinsic dependencies within the models, and improves memory efficiency of distributed ML. We demonstrate the efficacy of model-parallel algorithms implemented on STRADS versus popular implementations for topic modeling, matrix factorization, and Lasso.

Latent Dirichlet allocation (LDA) is a popular generative model of various objects such as texts and images, where an object is expressed as a mixture of latent topics. In this paper, we theoretically investigate variational Bayesian (VB) learning in LDA. More specifically, we analytically derive the leading term of the VB free energy under an asymptotic setup, and show that there exist transition thresholds in Dirichlet hyperparameters around which the sparsity-inducing behavior drastically changes. Then we further theoretically reveal the notable phenomenon that VB tends to induce weaker sparsity than MAP in the LDA model, which is opposed to other models. We experimentally demonstrate the practical validity of our asymptotic theory on real-world Last.FM music data.

Extracting 3D shape of deforming objects in monocular videos, a task known as non-rigid structure-from-motion (NRSfM), has so far been studied only on synthetic datasets and controlled environments. Typically, the objects to reconstruct are pre-segmented, they exhibit limited rotations and occlusions, or full-length trajectories are assumed. In order to integrate NRSfM into current video analysis pipelines, one needs to consider as input realistic -thus incomplete- tracking, and perform spatio-temporal grouping to segment the objects from their surroundings. Furthermore, NRSfM needs to be robust to noise in both segmentation and tracking, e.g., drifting, segmentation ``leaking'', optical flow ``bleeding'' etc. In this paper, we make a first attempt towards this goal, and propose a method that combines dense optical flow tracking, motion trajectory clustering and NRSfM for 3D reconstruction of objects in videos. For each trajectory cluster, we compute multiple reconstructions by minimizing the reprojection error and the rank of the 3D shape under different rank bounds of the trajectory matrix. We show that dense 3D shape is extracted and trajectories are completed across occlusions and low textured regions, even under mild relative motion between the object and the camera. We achieve competitive results on a public NRSfM benchmark while using fixed parameters across all sequences and handling incomplete trajectories, in contrast to existing approaches. We further test our approach on popular video segmentation datasets. To the best of our knowledge, our method is the first to extract dense object models from realistic videos, such as those found in Youtube or Hollywood movies, without object-specific priors.

In this paper, we propose an efficient semidefinite programming (SDP) approach to worst-case linear discriminant analysis (WLDA). Compared with the traditional LDA, WLDA considers the dimensionality reduction problem from the worst-case viewpoint, which is in general more robust for classification. However, the original problem of WLDA is non-convex and difficult to optimize. In this paper, we reformulate the optimization problem of WLDA into a sequence of semidefinite feasibility problems. To efficiently solve the semidefinite feasibility problems, we design a new scalable optimization method with quasi-Newton methods and eigen-decomposition being the core components. The proposed method is orders of magnitude faster than standard interior-point based SDP solvers. Experiments on a variety of classification problems demonstrate that our approach achieves better performance than standard LDA. Our method is also much faster and more scalable than standard interior-point SDP solvers based WLDA. The computational complexity for an SDP with $m$ constraints and matrices of size $d$ by $d$ is roughly reduced from $\mathcal{O}(m^3+md^3+m^2d^2)$ to $\mathcal{O}(d^3)$ ($m>d$ in our case).

One of the goals of any scientific study is to identify regularities in obs ervations and classify them into possibly separate and simpler structures or c ategories. These categories can in turn be used to make inferences on the obj ects of interest. The major advantage of this approach is that one breaks down a co mplicated reality into a collection of simpler structures. In a similar way, in patte rn recognition one is concern with discovery of regularities in data but t hrough use of computer algorithms which can be used to classify the data int o different categories [Bis06]. Independent of ones point of view, any such ana lysis must start with definition of the categories. If one has sufficient informa tion about the categories and their members, it is an easy task to establish a precis e definition. However, for most real life situations this is not the case and the no tion of category cannot be precisely defined. Under such conditions a fru itful approach is to consider a category as collection of objects which are likely to sh are the same properties.

We present a case-based approach to character identification in natural language text in the context of our Voz system. Voz first extracts entities from the text, and for each one of them, computes a feature-vector using both linguistic information and external knowledge. We propose a new similarity measure called Continuous Jaccard that exploits those feature-vectors to compute the similarity between a given entity and those in the case-base, and thus determine which entities are characters or not. We evaluate our approach by comparing it with different similarity measures and feature sets. Results show an identification accuracy of up to 93.49%, significantly higher than recent related work.

The increased use of large corpora in narrative research has created new opportunities for empirical research and intelligent narrative technologies. To best exploit the value of these corpora, several research groups are eschewing complex discourse analysis techniques in favor of high-level minimalist narrative annotation schemes that can be quickly applied, achieve high inter-rater agreement, and are amenable to automation using machine-learning techniques. In this paper we compare different annotation schemes that have been employed by two groups of researchers to annotate large corpora of narrative text. Using a dual-annotation methodology, we investigate the correlation between narrative clauses distinguished by their structural role (orientation, action, evaluation), their subjectivity, and their narrative level within the discourse. We find that each simple narrative annotation scheme captures a structurally distinct characteristic of real-world narratives, and each combination of labels is evident in a corpus of 19 weblog narratives (951 narrative clauses). We discuss several potential applications of minimalist narrative annotation schemes, noting the combination of label across these two annotation schemes that best support each task.