Algorithms frequently manage online advertising markets, aligning advertisements with article topics. Our work investigates how users perceive the relevance of ads to articles when ads are placed using different keyword extraction algorithms, including Large Language Models (LLMs), and how transparency about the placement procedure influences these perceptions and behavioral intentions. We conducted an online user experiment (N = 498) where ads are matched with news articles using the keyword extraction methods TF-IDF, KeyBERT, and DeepSeek. Results indicate that lightweight methods can match advanced LLMs in delivering high user-perceived ad-article relevance, which in turn fosters click and purchase intentions. However, providing explanations for the ad-article placements by displaying extracted keywords reduces ad interest and thereby weakens behavioral intentions, while simultaneously increasing perceived relevance and moderating algorithm effects. These findings highlight the complex impact of transparency-increasing explanations and suggest that algorithmic precision metrics must be complemented by user perception and intention measures.

We present VOICE, a novel approach to science communication that connects large language models' conversational capabilities with interactive exploratory visualization. VOICE introduces several innovative technical contributions that drive our conversational visualization framework. Based on the collected design requirements, we introduce a two-layer agent architecture that can perform task assignment, instruction extraction, and coherent content generation. We employ fine-tuning and prompt engineering techniques to tailor agents' performance to their specific roles and accurately respond to user queries. Our interactive text-to-visualization method generates a flythrough sequence matching the content explanation. In addition, natural language interaction provides capabilities to navigate and manipulate 3D models in real-time. The VOICE framework can receive arbitrary voice commands from the user and respond verbally, tightly coupled with a corresponding visual representation, with low latency and high accuracy. We demonstrate the effectiveness of our approach by implementing a proof-of-concept prototype and applying it to the molecular visualization domain: analyzing three 3D molecular models with multiscale and multi-instance attributes. Finally, we conduct a comprehensive evaluation of the system, including quantitative and qualitative analyses on our collected dataset, along with a detailed public user study and expert interviews. The results confirm that our framework and prototype effectively meet the design requirements and cater to the needs of diverse target users.

Dietze's conjecture concerns the problem of equipping a tree automaton M with weights to make it probabilistic, in such a way that the resulting automaton N predicts a given corpus C as accurately as possible. The conjecture states that the accuracy cannot increase if the states in M are merged with respect to an equivalence relation ∼ so that the result is a smaller automaton M∼. Put differently, merging states can never improve predictions. This is under the assumption that both M and M∼ are bottom-up deterministic and accept every tree in C. We prove that the conjecture holds, using a construction that turns any probabilistic version N∼ of M∼ into a probabilistic version N of M, such that N assigns at least as great a weight to each tree in C as N∼ does.

Abstract Graph-based semantic representations are popular in natural language processing, where it is often convenient to model linguistic concepts as nodes and relations as edges between them. Several attempts have been made to find a generative device that is sufficiently powerful to describe languages of semantic graphs, while at the same allowing efficient parsing. We contribute to this line of work by introducing graph extension grammar, a variant of the contextual hyperedge replacement grammars proposed by Hoffmann et al. Contextual hyperedge replacement can generate graphs with non-structural reentrancies, a type of node-sharing that is very common in formalisms such as abstract meaning representation, but that context-free types of graph grammars cannot model. To provide our formalism with a way to place reentrancies in a linguistically meaningful way, we endow rules with logical formulas in counting monadic second-order logic. We then present a parsing algorithm and show as our main result that this algorithm runs in polynomial time on graph languages generated by a subclass of our grammars, the so-called local graph extension grammars.

We introduce a fold operation that realises a tree-to-graph transduction by merging selected nodes in the input tree to form a possibly cyclic output graph. The work is motivated by the increasing use of graph-based representations in semantic parsing. We show that a suitable class of graphs languages can be generated by applying the fold operation to regular unranked tree languages. We investigate two versions of the fold operation, one that preserves a depth-first ordering between the edges, and one that does not. Finally, we demonstrate that the time complexity for the associated non-uniform membership problem is solvable in polynomial time for the order-preserving version, and NP-complete for the order-cancelling one.

Abstract We show that a previously proposed algorithm for the N-best trees problem can be made more efficient by changing how it arranges and explores the search space. Given an integer N and a weighted tree automaton (wta) M over the tropical semiring, the algorithm computes N trees of minimal weight with respect to M. Compared with the original algorithm, the modifications increase the laziness of the evaluation strategy, which makes the new algorithm asymptotically more efficient than its predecessor. The algorithm is implemented in the software Betty, and compared to the state-of-the-art algorithm for extracting the N best runs, implemented in the software toolkit Tiburon. The data sets used in the experiments are wtas resulting from real-world natural language processing tasks, as well as artificially created wtas with varying degrees of nondeterminism. We find that Betty outperforms Tiburon on all tested data sets with respect to running time, while Tiburon seems to be the more memory-efficient choice.

There is a growing consensus that surface form alone does not enable models to learn meaning and gain language understanding.This warrants an interest in hybrid systems that combine the strengths of neural and symbolic methods.We favour triadic systems consisting of neural networks, knowledge bases, and inference engines.The network provides perception, that is, the interface between the system and its environment.The knowledge base provides explicit memory and thus immediate access to established facts.Finally, inference capabilities are provided by the inference engine which reflects on the perception, supported by memory, to reason and discover new facts.In this work, we probe six popular language models for semantic relations and outline a future line of research to study how the constituent subsystems can be jointly realised and integrated.

Abstract We present a minimization algorithm for non-deterministic finite state automata that finds and merges bisimulation-equivalent states. The bisimulation relation is computed through partition aggregation, in contrast to existing algorithms that use partition refinement. The algorithm simultaneously generalises and simplifies an earlier one by Watson and Daciuk for deterministic devices. We show the algorithm to be correct and run in time $$ O \left( n^2 r^2 \left| \varSigma \right| \right) $$ O n 2 r 2 Σ , where n is the number of states of the input automaton $$M$$ M , r is the maximal out-degree in the transition graph for any combination of state and input symbol, and $$\left| \varSigma \right| $$ Σ is the size of the input alphabet. The algorithm has a higher time complexity than derivatives of Hopcroft’s partition-refinement algorithm, but represents a promising new solution approach that preserves language equivalence throughout the computation process. Furthermore, since the algorithm essentially computes the maximal model of a logical formula derived from $$M$$ M , optimisation techniques from the field of model checking become applicable.

We develop a finite-state transducer for translating unranked trees into general graphs. This work is motivated by recent progress in semantic parsing for natural language, where sentences are first mapped into tree-shaped syntactic representations, and then these trees are translated into graph semantic representations. We investigate formal properties of our tree-to-graph transducers and develop a polynomial time algorithm for translating a weighted language of input trees into a packed representation, from which best-score graphs can be efficiently recovered.

Semantic embeddings have advanced the state of the art for countless natural language processing tasks, and various extensions to multimodal domains, such as visual-semantic embeddings, have been proposed.While the power of visual-semantic embeddings comes from the distillation and enrichment of information through machine learning, their inner workings are poorly understood and there is a shortage of analysis tools.To address this problem, we generalize the notion of probing tasks to the visual-semantic case.To this end, we (i) discuss the formalization of probing tasks for embeddings of image-caption pairs, (ii) define three concrete probing tasks within our general framework, (iii) train classifiers to probe for those properties, and (iv) compare various state-of-the-art embeddings under the lens of the proposed probing tasks.Our experiments reveal an up to 12% increase in accuracy on visual-semantic embeddings compared to the corresponding unimodal embeddings, which suggest that the text and image dimensions represented in the former do complement each other.

Abstract The efficacy of syntactic features for topic-independent authorship attribution is evaluated, taking a feature set of frequencies of words and punctuation marks as baseline. The features are ‘deep’ in the sense that they are derived by parsing the subject texts, in contrast to ‘shallow’ syntactic features for which a part-of-speech analysis is enough. The experiments are made on two corpora of online texts and one corpus of novels written around the year 1900. The classification tasks include classical closed-world authorship attribution, identification of separate texts among the works of one author, and cross-topic authorship attribution. In the first tasks, the feature sets were fairly evenly matched, but for the last task, the syntax-based feature set outperformed the baseline feature set. These results suggest that, compared to lexical features, syntactic features are more robust to changes in topic.

Abstract We present a randomised inference algorithm for regular tree languages. The algorithm takes as input two disjoint finite nonempty sets of trees 𝒫 and 𝒩 and outputs a nondeterministic finite tree automaton that accepts every tree in 𝒫 and rejects every tree in 𝒩. The output automaton typically represents a nontrivial generalisation of the examples given in 𝒫 and 𝒩. To obtain compact output automata, we use a heuristics similar to bisimulation minimisation. The algorithm has time complexity of $\ordo{\negsize \cdot \possize^2}$ , where n𝒩 and n𝒫 are the size of 𝒩 and 𝒫, respectively. Experiments are conducted on a prototype implementation, and the empirical results appear to second the theoretical results.

Describing social positions and roles is an important topic within social network analysis. One approach is to compute a suitable equivalence relation on the nodes of the target network. One relation that is often used for this purpose is regular equivalence, or bisimulation, as it is known within the field of computer science. In this paper we consider a relation from computer science called simulation relation. Simulation creates a partial order on the set of actors in a network and we can use this order to identify actors that have characteristic properties. The simulation relation can also be used to compute simulation equivalence which is a less restrictive equivalence relation than regular equivalence but is still computable in polynomial time. This paper primarily considers weighted directed networks and we present definitions of both weighted simulation equivalence and weighted regular equivalence. Weighted networks can be used to model a number of network domains, including information flow, trust propagation, and communication channels. Many of these domains have applications within homeland security and in the military, where one wants to survey and elicit key roles within an organization. Identifying social positions can be difficult when the target organization lacks a formal structure or is partially hidden.

Several models of automata are available that operate unranked trees. Two well-known examples are the stepwise unranked tree automaton (suta) and the parallel unranked tree automaton (puta). By adding a weight, taken from some semiring, to every transition we generalise these two qualitative automata models to quantitative models, thereby obtaining weighted stepwise unranked tree automata (wsuta) and weighted parallel unranked tree automata (wputa); the qualitative automata models are reobtained by choosing the BOOLEAN semiring. The weighted versions have applications in natural language processing, XML-based data management and quantitative information retrieval. We address the minimisation problem of wsuta and wputa by using (forward and backward) bisimulations and we prove the following results: (1) for every wsuta an equivalent forward (resp. backward) bisimulation minimal wsuta can be computed in time O(mn) where n is the number of states and m is the number of transitions of the given wsuta; (2) the same result is proved for wputa instead of wsuta; (3) if the semiring is additive cancellative or the BOOLEAN semiring, then the bound can be improved to O(mlog n) for both wsuta and wputa; (4) for every deterministic puta we can compute a minimal equivalent deterministic puta in time O(mlog n); (5) the automata models wsuta, wputa, and weighted unranked tree automaton have the same computational power.