Large language models (LLMs) have shown promising performance across diverse domains. Many practical applications of LLMs, such as code completion and structured data extraction, require adherence to syntactic constraints specified by a formal language. Yet, due to their probabilistic nature, LLM output is not guaranteed to adhere to such formal languages. Prior work has proposed constrained decoding as a means to restrict LLM generation to particular formal languages. However, existing works are not applicable to the emerging paradigm of diffusion LLMs, when used in practical scenarios such as the generation of formally correct C++ or JSON output. In this paper we address this challenge and present the first constrained decoding method for diffusion models, one that can handle formal languages captured by context-free grammars. We begin by reducing constrained decoding to the more general additive infilling problem, which asks whether a partial output can be completed to a valid word in the target language. This problem also naturally subsumes the previously unaddressed multi-region infilling constrained decoding. We then reduce this problem to the task of deciding whether the intersection of the target language and a regular language is empty and present an efficient algorithm to solve it for context-free languages. Empirical results on various applications, such as C++ code infilling and structured data extraction in JSON, demonstrate that our method achieves near-perfect syntactic correctness while consistently preserving or improving functional correctness. Importantly, our efficiency optimizations ensure that the computational overhead remains practical.
Muzhaffar Hazman, Minh-Khoi Pham, Shweta Soundararajan
et al.
Prompt engineering has proven to be a crucial step in leveraging pretrained large language models (LLMs) in solving various real-world tasks. Numerous solutions have been proposed that seek to automate prompt engineering by using the model itself to edit prompts. However, the majority of state-of-the-art approaches are evaluated on tasks that require minimal prompt templates and on very large and highly capable LLMs. In contrast, solving complex tasks that require detailed information to be included in the prompt increases the amount of text that needs to be optimised. Furthermore, smaller models have been shown to be more sensitive to prompt design. To address these challenges, we propose an evolutionary search approach to automated discrete prompt optimisation consisting of two phases. In the first phase, grammar-guided genetic programming is invoked to synthesise prompt-creating programmes by searching the space of programmes populated by function compositions of syntactic, dictionary-based and LLM-based prompt-editing functions. In the second phase, local search is applied to explore the neighbourhoods of best-performing programmes in an attempt to further fine-tune their performance. Our approach outperforms three state-of-the-art prompt optimisation approaches, PromptWizard, OPRO, and RL-Prompt, on three relatively small general-purpose LLMs in four domain-specific challenging tasks. We also illustrate several examples where these benchmark methods suffer relatively severe performance degradation, while our approach improves performance in almost all task-model combinations, only incurring minimal degradation when it does not.
Arnav Attri, Anuj Attri, Pushpak Bhattacharyya
et al.
Product recommendations inherently involve comparisons, yet traditional opinion summarization often fails to provide holistic comparative insights. We propose the novel task of generating Query-Focused Comparative Explainable Summaries (QF-CES) using Multi-Source Opinion Summarization (M-OS). To address the lack of query-focused recommendation datasets, we introduce MS-Q2P, comprising 7,500 queries mapped to 22,500 recommended products with metadata. We leverage Large Language Models (LLMs) to generate tabular comparative summaries with query-specific explanations. Our approach is personalized, privacy-preserving, recommendation engine-agnostic, and category-agnostic. M-OS as an intermediate step reduces inference latency approximately by 40% compared to the direct input approach (DIA), which processes raw data directly. We evaluate open-source and proprietary LLMs for generating and assessing QF-CES. Extensive evaluations using QF-CES-PROMPT across 5 dimensions (clarity, faithfulness, informativeness, format adherence, and query relevance) showed an average Spearman correlation of 0.74 with human judgments, indicating its potential for QF-CES evaluation.
Romain Lopez, Jan-Christian Huetter, Ehsan Hajiramezanali
et al.
Deep Generative Models (DGMs) are versatile tools for learning data representations while adequately incorporating domain knowledge such as the specification of conditional probability distributions. Recently proposed DGMs tackle the important task of comparing data sets from different sources. One such example is the setting of contrastive analysis that focuses on describing patterns that are enriched in a target data set compared to a background data set. The practical deployment of those models often assumes that DGMs naturally infer interpretable and modular latent representations, which is known to be an issue in practice. Consequently, existing methods often rely on ad-hoc regularization schemes, although without any theoretical grounding. Here, we propose a theory of identifiability for comparative DGMs by extending recent advances in the field of non-linear independent component analysis. We show that, while these models lack identifiability across a general class of mixing functions, they surprisingly become identifiable when the mixing function is piece-wise affine (e.g., parameterized by a ReLU neural network). We also investigate the impact of model misspecification, and empirically show that previously proposed regularization techniques for fitting comparative DGMs help with identifiability when the number of latent variables is not known in advance. Finally, we introduce a novel methodology for fitting comparative DGMs that improves the treatment of multiple data sources via multi-objective optimization and that helps adjust the hyperparameter for the regularization in an interpretable manner, using constrained optimization. We empirically validate our theory and new methodology using simulated data as well as a recent data set of genetic perturbations in cells profiled via single-cell RNA sequencing.
Fernando Gabriela Garcia, Spencer Burns, Harrison Fuller
In this paper, we introduce ChatCite, a novel method leveraging large language models (LLMs) for generating comparative literature summaries. The ability to summarize research papers with a focus on key comparisons between studies is an essential task in academic research. Existing summarization models, while effective at generating concise summaries, fail to provide deep comparative insights. ChatCite addresses this limitation by incorporating a multi-step reasoning mechanism that extracts critical elements from papers, incrementally builds a comparative summary, and refines the output through a reflective memory process. We evaluate ChatCite on a custom dataset, CompLit-LongContext, consisting of 1000 research papers with annotated comparative summaries. Experimental results show that ChatCite outperforms several baseline methods, including GPT-4, BART, T5, and CoT, across various automatic evaluation metrics such as ROUGE and the newly proposed G-Score. Human evaluation further confirms that ChatCite generates more coherent, insightful, and fluent summaries compared to these baseline models. Our method provides a significant advancement in automatic literature review generation, offering researchers a powerful tool for efficiently comparing and synthesizing scientific research.
Comparative Judgement is an assessment method where item ratings are estimated based on rankings of subsets of the items. These rankings are typically pairwise, with ratings taken to be the estimated parameters from fitting a Bradley-Terry model. Likelihood penalization is often employed. Adaptive scheduling of the comparisons can increase the efficiency of the assessment. We show that the most commonly used penalty is not the best-performing penalty under adaptive scheduling and can lead to substantial bias in parameter estimates. We demonstrate this using simulated and real data and provide a theoretical explanation for the relative performance of the penalties considered. Further, we propose a superior approach based on bootstrapping. It is shown to produce better parameter estimates for adaptive schedules and to be robust to variations in underlying strength distributions and initial penalization method.
The twentieth century witnessed an abundant number of traumatic events related to dark history, like exiles and repressions by the Soviet regime in Lithuania in 1940-1953. In a single week of June 1941, the Soviets exiled 2% of the entire Lithuania’s population, while the total number reached nearly 14%. At the time, when it was allowed to speak about the unspeakable events of travelling to and surviving imprisonment in different concentration camps, numerous important works of various genres were published.
The first historical novel in English - Between Shades of Gray - was written in 2011 by Ruta Sepetys, the daughter of a Lithuanian refugee. The novel was translated into 30 languages. In 2018 Marius A. Markevicius adapted the novel into a film titled “Ashes in the Snow.”
The aim of the research is to discuss trauma and its reflection in literature and cinema, focusing on translation as screen adaptation. The novelty of the paper lies in the topic of (re)focalization when dealing with screen adaptation in relation to collective or personal traumas embodied in literary works. The concepts of conventional translation and adaptational translation by Henrik Gotlieb (2017: 52) are also discussed. The analysis of trauma is based on Cathy Caruth’s ideas who defines traumatic memories as non-verbal, so filmmakers have to find a way to express trauma when language becomes inaccessible and inadequate (Caruth 1996: 3-6).
Gerard Genette’s three types of focalization, - zero, internal and external, - as well as visible and invisible narrator in the story, offer a new approach to the study of audiovisual translation from the perspective of screen adaptation according to the external and internal position of the focalizer in the narrative: perceptual, psychological and ideological.
Questo articolo analizza la rappresentazione del lavoro considerando la sua profonda relazione con le questioni ambientali e il pensiero ecologico. Grazie a un'analisi ecopoetica di Strada Provinciale Tre di Vinci (2007), questa ricerca si propone di contribuire al dibattito critico-letterario intorno al lavoro mettendo a fuoco diversi aspetti solitamente non considerati: da un lato, l'interrelazione tra l'atto del lavorare e i cambiamenti ambientali, dall'altro, il potenziale potere delle immagini ecologiche legate al lavoro. Il primo aspetto sarà mostrato osservando gli effetti che la produzione capitalista ha sul corpo-lavoro e sul corpo-terra. Il secondo sarà esaminato attraverso tre diversi processi narrativi che contraddistinguono il romanzo di Vinci: la camminata immersiva, i ritmi naturali e antropici, gli aspetti sociali e spaziali delle pratiche mobili. Il quadro teorico combinerà la critica tematica del lavoro con gli studi sulle Mobilità e l'Ecopoetica.
Geography. Anthropology. Recreation, Language. Linguistic theory. Comparative grammar
Self-access language centres (SALCs) utilise self-directed learning, which allows learners to set their own learning goals and make their own learning choices. While this approach can benefit language learners, it can also be intimidating for those not used to autonomous learning. To help learners transition from teacher-directed to self-directed learning, various interventions such as coaching, counselling, mentoring, and advising can be used. These interventions can be effective in helping learners become more self-directed, but it is important to review and consider the learner’s expectations when implementing them. This paper analyses data from a one-on-one coaching intervention between an English-speaking coach and five Japanese participants using English as a second language to explore the factors contributing to the intervention session’s perceived success.
Andreas Orthey, Constantinos Chamzas, Lydia E. Kavraki
Sampling-based motion planning is one of the fundamental paradigms to generate robot motions, and a cornerstone of robotics research. This comparative review provides an up-to-date guideline and reference manual for the use of sampling-based motion planning algorithms. This includes a history of motion planning, an overview about the most successful planners, and a discussion on their properties. It is also shown how planners can handle special cases and how extensions of motion planning can be accommodated. To put sampling-based motion planning into a larger context, a discussion of alternative motion generation frameworks is presented which highlights their respective differences to sampling-based motion planning. Finally, a set of sampling-based motion planners are compared on 24 challenging planning problems. This evaluation gives insights into which planners perform well in which situations and where future research would be required. This comparative review thereby provides not only a useful reference manual for researchers in the field, but also a guideline for practitioners to make informed algorithmic decisions.
Morse complexes and Morse-Smale complexes are topological descriptors popular in topology-based visualization. Comparing these complexes plays an important role in their applications in feature correspondences, feature tracking, symmetry detection, and uncertainty visualization. Leveraging recent advances in optimal transport, we apply a class of optimal transport distances to the comparative analysis of Morse complexes. Contrasting with existing comparative measures, such distances are easy and efficient to compute, and naturally provide structural matching between Morse complexes. We perform an experimental study involving scientific simulation datasets and discuss the effectiveness of these distances as comparative measures for Morse complexes. We also provide an initial guideline for choosing the optimal transport distances under various data assumptions.
Antonio Artuñedo, Marcos Moreno-Gonzalez, Jorge Villagra
The selection of an appropriate control strategy is essential for ensuring safe operation in autonomous driving. While numerous control strategies have been developed for specific driving scenarios, a comprehensive comparative assessment of their performance using the same tuning methodology is lacking in the literature. This paper addresses this gap by presenting a systematic evaluation of state-of-the-art model-free and model-based control strategies. The objective is to evaluate and contrast the performance of these controllers across a wide range of driving scenarios, reflecting the diverse needs of autonomous vehicles. To facilitate the comparative analysis, a comprehensive set of performance metrics is selected, encompassing accuracy, robustness, and comfort. The contributions of this research include the design of a systematic tuning methodology, the use of two novel metrics for stability and comfort comparisons and the evaluation through extensive simulations and real tests in an experimental instrumented vehicle over a wide range of trajectories.
A filoop is a generic immersion of a circle in a closed oriented surface, whose complement is a disjoint union of discs, considered up to orientation preserving diffeomorphisms. It gives rise to a chord diagram C which has an interlace graph G, called a chordiagraph. For a graph G with even degrees, we compute a quantity mg(G) which yields, for every chord diagram $C$ with interlace graph G, the minimal genus of filoops with chord diagram C. If mg(G)=0 then C admits exactly two framings of genus 0, corresponding to spheriloops. After recalling the Cunningham factorisation of connected graphs, we describe a canonical factorisation of filoops into spheric sums followed by toric sums, for which the genus is additive. This is analogous to the factorisation of compact connected 3-manifolds along spheres and tori. We describe unambiguous context-sensitive grammars generating the set of all graphs and with mg(G)=0 and deduce stability properties with respect to spheric and toric factorisations. Similar results hold for chordiagraphs with mg(G) = 0 and their corresponding spheriloops.
Abstract Spoken language, as we have it, requires specific capacities—at its most basic advanced vocal control and complex social cognition. In humans, vocal control is the basis for speech, achieved through coordinated interactions of larynx activity and rapid changes in vocal tract configurations. Most likely, speech evolved in response to early humans perceiving reality in increasingly complex ways, to the effect that primate‐like signaling became unsustainable as a sole communication device. However, in what ways did and do humans see the world in more complex ways compared to other species? Although animal signals can refer to external events, in contrast to humans, they usually refer to the agents only, sometimes in compositional ways, but never together with patients. It may be difficult for animals to comprehend events as part of larger social scripts, with antecedent causes and future consequences, which are more typically tie the patient into the event. Human brain enlargement over the last million years probably has provided the cognitive resources to represent social interactions as part of bigger social scripts, which enabled humans to go beyond an agent‐focus to refer to agent–patient relations, the likely foundation for the evolution of grammar. This article is categorized under: Cognitive Biology > Evolutionary Roots of Cognition Linguistics > Evolution of Language Psychology > Comparative
Historia języka ukraińskiego w artykule jest rozpatrywana z punktu widzenia planowania językowego – koncepcje H. Klossa, J. Fishmana oraz W. Crofta. Procesy ausbauzacji i einbauzacji, szczegóły regionalne całego kontinuum dialektalnego zostały opisane z uwzględnieniem wpływów języków: rosyjskiego, polskiego, węgierskiego i rumuńskiego.
The discovery of neural architectures from simple building blocks is a long-standing goal of Neural Architecture Search (NAS). Hierarchical search spaces are a promising step towards this goal but lack a unifying search space design framework and typically only search over some limited aspect of architectures. In this work, we introduce a unifying search space design framework based on context-free grammars that can naturally and compactly generate expressive hierarchical search spaces that are 100s of orders of magnitude larger than common spaces from the literature. By enhancing and using their properties, we effectively enable search over the complete architecture and can foster regularity. Further, we propose an efficient hierarchical kernel design for a Bayesian Optimization search strategy to efficiently search over such huge spaces. We demonstrate the versatility of our search space design framework and show that our search strategy can be superior to existing NAS approaches. Code is available at https://github.com/automl/hierarchical_nas_construction.
We introduce a family of comparative plausibility logics over neighbourhood models, generalising Lewis' comparative plausibility operator over sphere models. We provide axiom systems for the logics, and prove their soundness and completeness with respect to the semantics. Then, we introduce two kinds of analytic proof systems for several logics in the family: a multi-premisses sequent calculus in the style of Lellmann and Pattinson, for which we prove cut admissibility, and a hypersequent calculus based on structured calculi for conditional logics by Girlando et al., tailored for countermodel construction over failed proof search. Our results constitute the first steps in the definition of a unified proof theoretical framework for logics equipped with a comparative plausibility operator.