arXiv subjects

Jakob Grue Simonsen

Publications and source records attributed to Jakob Grue Simonsen.

At least 37 records · Page 2Linked to original sources

arXiv2020

A Complete Characterization of Infinitely Repeated Two-Player Games having Computable Strategies with no Computable Best Response under Limit-of-Means Payoff

It is well-known that for infinitely repeated games, there are computable strategies that have best responses, but no computable best responses. These results were originally proved for either specific games (e.g., Prisoner's dilemma), or for classes of games satisfying certain conditions not known to be both necessary and sufficient. We derive a complete characterization in the form of simple necessary and sufficient conditions for the existence of a computable strategy without a computable best response under limit-of-means payoff. We further refine the characterization by requiring the strategy profiles to be Nash equilibria or subgame-perfect equilibria, and we show how the characterizations entail that it is efficiently decidable whether an infinitely repeated game has a computable strategy without a computable best response.

cs.GT↗

arXiv2020

Content-aware Neural Hashing for Cold-start Recommendation

Content-aware recommendation approaches are essential for providing meaningful recommendations for \textit{new} (i.e., \textit{cold-start}) items in a recommender system. We present a content-aware neural hashing-based collaborative filtering approach (NeuHash-CF), which generates binary hash codes for users and items, such that the highly efficient Hamming distance can be used for estimating user-item relevance. NeuHash-CF is modelled as an autoencoder architecture, consisting of two joint hashing components for generating user and item hash codes. Inspired from semantic hashing, the item hashing component generates a hash code directly from an item's content information (i.e., it generates cold-start and seen item hash codes in the same manner). This contrasts existing state-of-the-art models, which treat the two item cases separately. The user hash codes are generated directly based on user id, through learning a user embedding matrix. We show experimentally that NeuHash-CF significantly outperforms state-of-the-art baselines by up to 12\% NDCG and 13\% MRR in cold-start recommendation settings, and up to 4\% in both NDCG and MRR in standard settings where all items are present while training. Our approach uses 2-4x shorter hash codes, while obtaining the same or better performance compared to the state of the art, thus consequently also enabling a notable storage reduction.

cs.IR↗

arXiv2020

Generating Fact Checking Explanations

Most existing work on automated fact checking is concerned with predicting the veracity of claims based on metadata, social network spread, language used in claims, and, more recently, evidence supporting or denying claims. A crucial piece of the puzzle that is still missing is to understand how to automate the most elaborate part of the process -- generating justifications for verdicts on claims. This paper provides the first study of how these explanations can be generated automatically based on available claim context, and how this task can be modelled jointly with veracity prediction. Our results indicate that optimising both objectives at the same time, rather than training them separately, improves the performance of a fact checking system. The results of a manual evaluation further suggest that the informativeness, coverage and overall quality of the generated explanations are also improved in the multi-task model.

cs.CL↗

arXiv2019

MultiFC: A Real-World Multi-Domain Dataset for Evidence-Based Fact Checking of Claims

We contribute the largest publicly available dataset of naturally occurring factual claims for the purpose of automatic claim verification. It is collected from 26 fact checking websites in English, paired with textual sources and rich metadata, and labelled for veracity by human expert journalists. We present an in-depth analysis of the dataset, highlighting characteristics and challenges. Further, we present results for automatic veracity prediction, both with established baselines and with a novel method for joint ranking of evidence pages and predicting veracity that outperforms all baselines. Significant performance increases are achieved by encoding evidence, and by modelling metadata. Our best-performing model achieves a Macro F1 of 49.2%, showing that this is a challenging testbed for claim veracity prediction.

cs.CL↗

arXiv2019

Unsupervised Neural Generative Semantic Hashing

Fast similarity search is a key component in large-scale information retrieval, where semantic hashing has become a popular strategy for representing documents as binary hash codes. Recent advances in this area have been obtained through neural network based models: generative models trained by learning to reconstruct the original documents. We present a novel unsupervised generative semantic hashing approach, \textit{Ranking based Semantic Hashing} (RBSH) that consists of both a variational and a ranking based component. Similarly to variational autoencoders, the variational component is trained to reconstruct the original document conditioned on its generated hash code, and as in prior work, it only considers documents individually. The ranking component solves this limitation by incorporating inter-document similarity into the hash code generation, modelling document ranking through a hinge loss. To circumvent the need for labelled data to compute the hinge loss, we use a weak labeller and thus keep the approach fully unsupervised. Extensive experimental evaluation on four publicly available datasets against traditional baselines and recent state-of-the-art methods for semantic hashing shows that RBSH significantly outperforms all other methods across all evaluated hash code lengths. In fact, RBSH hash codes are able to perform similarly to state-of-the-art hash codes while using 2-4x fewer bits.

cs.IR↗

arXiv2019

Contextually Propagated Term Weights for Document Representation

Word embeddings predict a word from its neighbours by learning small, dense embedding vectors. In practice, this prediction corresponds to a semantic score given to the predicted word (or term weight). We present a novel model that, given a target word, redistributes part of that word's weight (that has been computed with word embeddings) across words occurring in similar contexts as the target word. Thus, our model aims to simulate how semantic meaning is shared by words occurring in similar contexts, which is incorporated into bag-of-words document representations. Experimental evaluation in an unsupervised setting against 8 state of the art baselines shows that our model yields the best micro and macro F1 scores across datasets of increasing difficulty.

cs.IR↗

arXiv2019

Neural Speed Reading with Structural-Jump-LSTM

Recurrent neural networks (RNNs) can model natural language by sequentially 'reading' input tokens and outputting a distributed representation of each token. Due to the sequential nature of RNNs, inference time is linearly dependent on the input length, and all inputs are read regardless of their importance. Efforts to speed up this inference, known as 'neural speed reading', either ignore or skim over part of the input. We present Structural-Jump-LSTM: the first neural speed reading model to both skip and jump text during inference. The model consists of a standard LSTM and two agents: one capable of skipping single words when reading, and one capable of exploiting punctuation structure (sub-sentence separators (,:), sentence end symbols (.!?), or end of text markers) to jump ahead after reading a word. A comprehensive experimental evaluation of our model against all five state-of-the-art neural reading models shows that Structural-Jump-LSTM achieves the best overall floating point operations (FLOP) reduction (hence is faster), while keeping the same accuracy or even improving it compared to a vanilla LSTM that reads the whole text.

cs.CL↗

arXiv2019

Neural Check-Worthiness Ranking with Weak Supervision: Finding Sentences for Fact-Checking

Automatic fact-checking systems detect misinformation, such as fake news, by (i) selecting check-worthy sentences for fact-checking, (ii) gathering related information to the sentences, and (iii) inferring the factuality of the sentences. Most prior research on (i) uses hand-crafted features to select check-worthy sentences, and does not explicitly account for the recent finding that the top weighted terms in both check-worthy and non-check-worthy sentences are actually overlapping [15]. Motivated by this, we present a neural check-worthiness sentence ranking model that represents each word in a sentence by \textit{both} its embedding (aiming to capture its semantics) and its syntactic dependencies (aiming to capture its role in modifying the semantics of other terms in the sentence). Our model is an end-to-end trainable neural network for check-worthiness ranking, which is trained on large amounts of unlabelled data through weak supervision. Thorough experimental evaluation against state of the art baselines, with and without weak supervision, shows our model to be superior at all times (+13% in MAP and +28% at various Precision cut-offs from the best baseline with statistical significance). Empirical analysis of the use of weak supervision, word embedding pretraining on domain-specific data, and the use of syntactic dependencies of our model reveals that check-worthy sentences contain notably more identical syntactic dependencies than non-check-worthy sentences.

cs.IR↗

arXiv2019

Modelling Sequential Music Track Skips using a Multi-RNN Approach

Modelling sequential music skips provides streaming companies the ability to better understand the needs of the user base, resulting in a better user experience by reducing the need to manually skip certain music tracks. This paper describes the solution of the University of Copenhagen DIKU-IR team in the 'Spotify Sequential Skip Prediction Challenge', where the task was to predict the skip behaviour of the second half in a music listening session conditioned on the first half. We model this task using a Multi-RNN approach consisting of two distinct stacked recurrent neural networks, where one network focuses on encoding the first half of the session and the other network focuses on utilizing the encoding to make sequential skip predictions. The encoder network is initialized by a learned session-wide music encoding, and both of them utilize a learned track embedding. Our final model consists of a majority voted ensemble of individually trained models, and ranked 2nd out of 45 participating teams in the competition with a mean average accuracy of 0.641 and an accuracy on the first skip prediction of 0.807. Our code is released at https://github.com/Varyn/WSDM-challenge-2019-spotify.

cs.IR↗

arXiv2017

Higher-order Cons-free Interpreters

Constructor rewriting systems are said to be cons-free if any constructor term occurring in the rhs of a rule must be a subterm of the lhs of the rule. Roughly, such systems cannot build new data structures during their evaluation. In earlier work by several authors, (typed) cons-free systems have been used to characterise complexity classes such as polynomial or exponential time or space by varying the type orders, and the recursion forms allowed. This paper concerns the construction of interpreters for cons-free term rewriting. Due to their connection with proofs by diagonalisation, interpreters may be of use when studying separation results between complexity classes in implicit computational complexity theory. We are interested in interpreters of type order $k > 1$ that can interpret any term of strictly lower type order; while this gives us a well-known separation result E$^k$TIME $\subseteq$ E$^{k+1}$TIME, the hope is that more refined interpreters with syntactically limited constraints can be used to obtain a notion of faux diagonalisation and be used to attack open problems in complexity theory.

cs.LO↗

arXiv2017

Evaluation Measures for Relevance and Credibility in Ranked Lists

Recent discussions on alternative facts, fake news, and post truth politics have motivated research on creating technologies that allow people not only to access information, but also to assess the credibility of the information presented to them by information retrieval systems. Whereas technology is in place for filtering information according to relevance and/or credibility, no single measure currently exists for evaluating the accuracy or precision (and more generally effectiveness) of both the relevance and the credibility of retrieved results. One obvious way of doing so is to measure relevance and credibility effectiveness separately, and then consolidate the two measures into one. There at least two problems with such an approach: (I) it is not certain that the same criteria are applied to the evaluation of both relevance and credibility (and applying different criteria introduces bias to the evaluation); (II) many more and richer measures exist for assessing relevance effectiveness than for assessing credibility effectiveness (hence risking further bias). Motivated by the above, we present two novel types of evaluation measures that are designed to measure the effectiveness of both relevance and credibility in ranked lists of retrieval results. Experimental evaluation on a small human-annotated dataset (that we make freely available to the research community) shows that our measures are expressive and intuitive in their interpretation.

cs.IR↗

arXiv2017

Complexity Hierarchies and Higher-order Cons-free Term Rewriting

Constructor rewriting systems are said to be cons-free if, roughly, constructor terms in the right-hand sides of rules are subterms of the left-hand sides; the computational intuition is that rules cannot build new data structures. In programming language research, cons-free languages have been used to characterize hierarchies of computational complexity classes; in term rewriting, cons-free first-order TRSs have been used to characterize the class PTIME. We investigate cons-free higher-order term rewriting systems, the complexity classes they characterize, and how these depend on the type order of the systems. We prove that, for every K $\geq$ 1, left-linear cons-free systems with type order K characterize E$^K$TIME if unrestricted evaluation is used (i.e., the system does not have a fixed reduction strategy). The main difference with prior work in implicit complexity is that (i) our results hold for non-orthogonal term rewriting systems with no assumptions on reduction strategy, (ii) we consequently obtain much larger classes for each type order (E$^K$TIME versus EXP$^{K-1}$TIME), and (iii) results for cons-free term rewriting systems have previously only been obtained for K = 1, and with additional syntactic restrictions besides cons-freeness and left-linearity. Our results are among the first implicit characterizations of the hierarchy E = E$^1$TIME $\subsetneq$ E$^2$TIME $\subsetneq$ ... Our work confirms prior results that having full non-determinism (via overlapping rules) does not directly allow for characterization of non-deterministic complexity classes like NE. We also show that non-determinism makes the classes characterized highly sensitive to minor syntactic changes like admitting product types or non-left-linear rules.

cs.CC↗

arXiv2017

Computability in the Lattice of Equivalence Relations

We investigate computability in the lattice of equivalence relations on the natural numbers. We mostly investigate whether the subsets of appropriately defined subrecursive equivalence relations -for example the set of all polynomial-time decidable equivalence relations- form sublattices of the lattice.

cs.CC↗

arXiv2017

Chains, Antichains, and Complements in Infinite Partition Lattices

We consider the partition lattice $Π_κ$ on any set of transfinite cardinality $κ$ and properties of $Π_κ$ whose analogues do not hold for finite cardinalities. Assuming the Axiom of Choice we prove: (I) the cardinality of any maximal well-ordered chain is always exactly $κ$; (II) there are maximal chains in $Π_κ$ of cardinality $> κ$; (III) if, for every cardinal $λ< κ$, we have $2^λ < 2^κ$, there exists a maximal chain of cardinality $< 2^κ$ (but $\ge κ$) in $Π_{2^κ}$; (IV) every non-trivial maximal antichain in $Π_κ$ has cardinality between $κ$ and $2^κ$, and these bounds are realized. Moreover we can construct maximal antichains of cardinality $\max(κ, 2^λ)$ for any $λ\le κ$; (V) all cardinals of the form $κ^λ$ with $0 \le λ\le κ$ occur as the number of complements to some partition $\mathcal{P} \in Π_κ$, and only these cardinalities appear. Moreover, we give a direct formula for the number of complements to a given partition; (VI) Under the Generalized Continuum Hypothesis, the cardinalities of maximal chains, maximal antichains, and numbers of complements are fully determined, and we provide a complete characterization.

math.RA↗

arXiv2017

The Power of Non-Determinism in Higher-Order Implicit Complexity

We investigate the power of non-determinism in purely functional programming languages with higher-order types. Specifically, we consider cons-free programs of varying data orders, equipped with explicit non-deterministic choice. Cons-freeness roughly means that data constructors cannot occur in function bodies and all manipulation of storage space thus has to happen indirectly using the call stack. While cons-free programs have previously been used by several authors to characterise complexity classes, the work on non-deterministic programs has almost exclusively considered programs of data order 0. Previous work has shown that adding explicit non-determinism to cons-free programs taking data of order 0 does not increase expressivity; we prove that this - dramatically - is not the case for higher data orders: adding non-determinism to programs with data order at least 1 allows for a characterisation of the entire class of elementary-time decidable sets. Finally we show how, even with non-deterministic choice, the original hierarchy of characterisations is restored by imposing different restrictions.

cs.CC↗

arXiv2016

Adaptive Distributional Extensions to DFR Ranking

Divergence From Randomness (DFR) ranking models assume that informative terms are distributed in a corpus differently than non-informative terms. Different statistical models (e.g. Poisson, geometric) are used to model the distribution of non-informative terms, producing different DFR models. An informative term is then detected by measuring the divergence of its distribution from the distribution of non-informative terms. However, there is little empirical evidence that the distributions of non-informative terms used in DFR actually fit current datasets. Practically this risks providing a poor separation between informative and non-informative terms, thus compromising the discriminative power of the ranking model. We present a novel extension to DFR, which first detects the best-fitting distribution of non-informative terms in a collection, and then adapts the ranking computation to this best-fitting distribution. We call this model Adaptive Distributional Ranking (ADR) because it adapts the ranking to the statistics of the specific dataset being processed each time. Experiments on TREC data show ADR to outperform DFR models (and their extensions) and be comparable in performance to a query likelihood language model (LM).

cs.IR↗

arXiv2016

Exploiting the Bipartite Structure of Entity Grids for Document Coherence and Retrieval

Document coherence describes how much sense text makes in terms of its logical organisation and discourse flow. Even though coherence is a relatively difficult notion to quantify precisely, it can be approximated automatically. This type of coherence modelling is not only interesting in itself, but also useful for a number of other text processing tasks, including Information Retrieval (IR), where adjusting the ranking of documents according to both their relevance and their coherence has been shown to increase retrieval effectiveness [34,37]. The state of the art in unsupervised coherence modelling represents documents as bipartite graphs of sentences and discourse entities, and then projects these bipartite graphs into one-mode undirected graphs. However, one-mode projections may incur significant loss of the information present in the original bipartite structure. To address this we present three novel graph metrics that compute document coherence on the original bipartite graph of sentences and entities. Evaluation on standard settings shows that: (i) one of our coherence metrics beats the state of the art in terms of coherence accuracy; and (ii) all three of our coherence metrics improve retrieval effectiveness because, as closer analysis reveals, they capture aspects of document quality that go undetected by both keyword-based standard ranking and by spam filtering. This work contributes document coherence metrics that are theoretically principled, parameter-free, and useful to IR.

cs.IR↗

arXiv2016

Deep Learning Relevance: Creating Relevant Information (as Opposed to Retrieving it)

What if Information Retrieval (IR) systems did not just retrieve relevant information that is stored in their indices, but could also "understand" it and synthesise it into a single document? We present a preliminary study that makes a first step towards answering this question. Given a query, we train a Recurrent Neural Network (RNN) on existing relevant information to that query. We then use the RNN to "deep learn" a single, synthetic, and we assume, relevant document for that query. We design a crowdsourcing experiment to assess how relevant the "deep learned" document is, compared to existing relevant documents. Users are shown a query and four wordclouds (of three existing relevant documents and our deep learned synthetic document). The synthetic document is ranked on average most relevant of all.

cs.IR↗