Video summarization is a challenging problem with great application potential. Whereas prior approaches, largely unsupervised in nature, focus on sampling useful frames and assembling them as summaries, we consider video summarization as a supervised subset selection problem. Our idea is to teach the system to learn from human-created summaries how to select informative and diverse subsets, so as to best meet evaluation metrics derived from human-perceived quality. To this end, we propose the sequential determinantal point process (seqDPP), a probabilistic model for diverse sequential subset selection. Our novel seqDPP heeds the inherent sequential structures in video data, thus overcoming the deficiency of the standard DPP, which treats video frames as randomly permutable items. Meanwhile, seqDPP retains the power of modeling diverse subsets, essential for summarization. Our extensive results of summarizing videos from 3 datasets demonstrate the superior performance of our method, compared to not only existing unsupervised methods but also naive applications of the standard DPP model.

We propose a spiking network model capable of performing both approximate inference and learning for any hidden Markov model. The lower layer sensory neurons detect noisy measurements of hidden world states. The higher layer neurons with recurrent connections infer a posterior distribution over world states from spike trains generated by sensory neurons. We show how such a neuronal network with synaptic plasticity can implement a form of Bayesian inference similar to Monte Carlo methods such as particle filtering. Each spike in the population of inference neurons represents a sample of a particular hidden world state.

First provide a summary of the paper, and then address the following criteria: Quality, clarity, originality and significance. In this paper, authors analyze sparsity of the posterior parameters in LDA using a variational Bayesian algorithm. They derive an expression for the VB free energy which shows its asymptotic behaviour with respect to number of words (N), number of documents (M), vocabulary size (L) etc. Their results suggest that, for certain settings of L,M,N, the sparsity behaviour changes drastically at a particular hyper-parameter setting. These changes differ from those of MAP and partial-Bayes algorithms. The problem discussed in this paper is original, interesting, and is perhaps useful too.

First provide a summary of the paper, and then address the following criteria: Quality, clarity, originality and significance. This paper proposes the Latent Case Model (LCM), a Bayesian approach to clustering in which clusters are represented by a prototype (a specific sample from the data) and feature subspaces (a binary subset of the variables signifying those features that are relevant to the class). The approach is presented as being a Bayesian, trainable version of the Case-Based Reasoning approach popular in AI, and is motivated by the ways such models have proved highly effective in explaining human decision making. The generative model (Figure 1) represents each item as coming from a mixture of S clusters, where each cluster is represented by a prototype and subspace (as above) and a function \phi which generates features matching those of the prototype with high probability for features in the subspace, and uniform features outside it. The model is thus similar in functionality to LDA but quite different in terms of its representation.

Humans are able to rapidly learn new concepts and continuously integrate them among prior knowledge.

We thank all reviewers for thoughtful feedback! We reply separately to each reviewer. Reviewer #1: We would like to point out some of the paper's main contributions, not fully recognized in the review. Another example is our algorithm for sampling DAGs conditionally on a root-partition (Sections 3.4 Accordingly, our main innovations are algorithmic. We would like to correct that our algorithm for sampling DAGs is not "classical" (cf.

First provide a summary of the paper, and then address the following criteria: Quality, clarity, originality and significance. Overview: The paper proposes a framework for enforcing structure in Bayesian models via structured prior selection based on the maximum entropy principle. Although the optimal prior may not be tractable, the authors developed an approximation method using submodule optimization. Contructing priors with structured variables is an important topic, so this method should be able to make good impact. Quality The paper is technically sound.