This paper strives to find amidst a set of sentences the one best describing the content of a given image or video. Different from existing works, which rely on a joint subspace for their image and video caption retrieval, we propose to do so in a visual space exclusively. Apart from this conceptual novelty, we contribute Word2VisualVec, a deep neural network architecture that learns to predict...

Different mathematical techniques are being developed to reduce the dimensionality of data within large datasets, for robust retrieval of required information. Latent Semantic Analysis (LSA), a modified low rank approximation form of Vector Space Model, can be used for detecting underlying semantic relationships within text corpora. LSA performs a low-rank approximation on term-document matrix,...

The constant growth of digital information, facilitated by storage technologies, imposes new challenges for information processing tasks, and maintains the need of effective search mechanisms, oriented towards improving in precision but simultaneously capable of producing useful information in a short time. Hence, this paper presents a document representation to encode textual relations. This r...

Modeling textual or visual information with vector representations trained from large language or visual datasets has been successfully explored in recent years. However, tasks such as visual question answering require combining these vector representations with each other. Approaches to multimodal pooling include element-wise multiplication or addition, as well as concatenation of the visual a...

The main goal of FBK-irst participation at RTE-4 was to experiment the use of combined specialized entailment engines, each addressing a specific phenomena relevant to entailment. The approach is motivated since textual entailment is due to the combination of several linguistic phenomena which interact among them in a quite complex way. We were driven by the following two considerations: (i) de...

This paper presents a methodology for a quantitative and qualitative evaluation of Textual Entailment systems. We take advantage of the decomposition of Text Hypothesis pairs into monothematic pairs, i.e. pairs where only one linguistic phenomenon at a time is responsible for entailment judgment, and propose to run TE systems over such datasets. We show that several behaviours of a system can b...

Hypernymy, textual entailment, and image captioning can be seen as special cases of a single visual-semantic hierarchy over words, sentences, and images. In this paper we advocate for explicitly modeling the partial order structure of this hierarchy. Towards this goal, we introduce a general method for learning ordered representations, and show how it can be applied to a variety of tasks involv...

We present a novel text exploration model, which extends the scope of state-of-the-art technologies by moving from standard concept-based exploration to statement-based exploration. The proposed scheme utilizes the textual entailment relation between statements as the basis of the exploration process. A user of our system can explore the result space of a query by drilling down/up from one stat...

The Stanford Dependencies are a deep syntactic representation that are widely used for semantic tasks, like Recognizing Textual Entailment. But do they capture all of the semantic information a meaning representation ought to convey? This paper explores this question by investigating the feasibility of mapping Stanford dependency parses to Hobbsian Logical Form, a practical, event-theoretic sem...