Journal: Frontiers in Artificial Intelligence and Applications

Abbreviation

Front. artif. intell. appl.

Publisher

IOS Press

Journal Volumes

ISSN

0922-6389
1879-8314

Description

Filters

Reset filters

Search Results

Publications 1 - 9 of 9
  • RETRO-LI: Small-Scale Retrieval Augmented Generation Supporting Noisy Similarity Searches and Domain Shift Generalization
    Item type: Conference Paper
    Rashiti, Gentiana; Karunaratne, Geethan; Sachan, Mrinmaya; et al. (2024)
    Frontiers in Artificial Intelligence and Applications ~ ECAI 2024
    The retrieval augmented generation (RAG) system such as RETRO has been shown to improve language modeling capabilities and reduce toxicity and hallucinations by retrieving from a database of non-parametric memory containing trillions of entries. We introduce RETRO-LI that shows retrieval can also help using a small scale database, but it demands more accurate and better neighbors when searching in a smaller hence sparser nonparametric memory. This can be met by using a proper semantic similarity search. We further propose adding a regularization to the non-parametric memory for the first time: it significantly reduces perplexity when the neighbor search operations are noisy during inference, and it improves generalization when a domain shift occurs. We also show that the RETRO-LI’s non-parametric memory can potentially be implemented on analog in-memory computing hardware, exhibiting O(1) search time while causing noise in retrieving neighbors, with minimal (<1%) performance loss. Our code is available at: https://github.com/IBM/Retrieval-Enhanced-Transformer-Little
  • Regulations Expressed As Logical Models (REALM)
    Item type: Conference Paper
    Giblin, Christopher; Liu, Alice Y.; Müller, Samuel; et al. (2005)
    Frontiers in Artificial Intelligence and Applications ~ Legal knowledge and information systems
  • Uncertainty measure of pythagorean fuzzy sets
    Item type: Conference Paper
    Gao, Xiaozhuan; Pan, Lipeng; Deng, Yong (2021)
    Frontiers in Artificial Intelligence and Applications
    Pythagorean fuzzy sets (PFS) can better express and handle the uncertainty information and has the more lager representation space. Hence, the reasonable and effective method to measure the uncertainty of PFS can better analyze information. From the view of Dempster-Shafer evidence theory, hesitancy degree can include the two focal elements (member-ship, non-membership). Hence, considering the number of focal elements for hesitancy degree to measure uncertainty is important. In addition, the difference between membership and non-membership degree plays an essential role in uncertainty measure. From the above views, the paper proposed the new uncertainty measure. Based on the new uncertainty measure, cross entropy and divergence of PFS can be presented. In addition, some numerical examples are used to explain the proposed methods by comparing other methods. Finally, the proposed divergence can be used in pattern recognition to verify its effectiveness.
  • A Phoneme-Based Student Model for Adaptive Spelling Training
    Item type: Conference Paper
    Baschera, Gian-Marco; Gross, Markus (2009)
    Frontiers in Artificial Intelligence and Applications ~ Artificial intelligence in education : building learning systems that care: from knowledge representation to affective modelling
  • Knowing whether spatio-temporal analysis procedures are applicable to datasets
    Item type: Conference Paper
    Scheider, Simon; Tomko, Martin (2016)
    Frontiers in Artificial Intelligence and Applications ~ Formal Ontology in Information Systems
    How can data analysts identify spatio-temporal datasets that are suitable for their task? Answering this question is not only dependent on the aim of the analysis and the semantic contents of the data, but also on knowing whether the required data combinations and transformations, spatio-temporal analysis methods, charts and map visualizations are meaningfully applicable to the data. Operators need to assess whether they can meaningfully apply analytical operations to data to derive the information required. Answering this question in a general and computationally executable way is a crucial step on our way towards supporting data analysts and their research practice in e-Science. We propose an ontology design pattern for spatio-temporal information that enables to reason about the applicability of a number of fundamental classes of analyses in relation to given data, i.e., whether data sets can be compared, transformed, combined, and whether summary statistics can be applied to them. We demonstrate this ontology implemented in OWL through a set of corresponding SPARQL queries applied to meta-data of datasets from the AURIN portal.
  • Diffusion of Pedagogical Innovations as a Complex Adaptive Process - Agent-Based Modeling as Research Method
    Item type: Conference Paper
    Huang, Jun-Song; Kapur, Manu (2007)
    Frontiers in Artificial Intelligence and Applications ~ Supporting learning flow through integrative technologies
  • SUBER: An RL Environment with Simulated Human Behavior for Recommender Systems
    Item type: Conference Paper
    Corecco, Nathan; Piatti, Giorgio; Lanzendörfer, Luca A.; et al. (2024)
    Frontiers in Artificial Intelligence and Applications ~ ECAI 2024
    Reinforcement learning (RL) has gained popularity in the realm of recommender systems due to its ability to optimize long-term rewards and guide users in discovering relevant content. However, the successful implementation of RL in recommender systems is challenging because of several factors, including the limited availability of online data for training on-policy methods. This scarcity requires expensive human interaction for online model training. Furthermore, the development of effective evaluation frameworks that accurately reflect the quality of models remains a fundamental challenge in recommender systems. To address these challenges, we propose a comprehensive framework for synthetic environments that simulate human behavior by harnessing the capabilities of large language models (LLMs). We complement our framework with in-depth ablation studies and demonstrate its effectiveness with experiments on movie and book recommendations. Using LLMs as synthetic users, this work introduces a modular and novel framework to train RL-based recommender systems. The software, including the RL environment, is publicly available on unmapped: uri https://github.com/SUBER-Team/SUBER.
  • GraphFSA: A Finite State Automaton Framework for Algorithmic Learning on Graphs
    Item type: Conference Paper
    Grötschla, Florian; Mathys, Joël; Raun, Christoffer; et al. (2024)
    Frontiers in Artificial Intelligence and Applications ~ ECAI 2024
    Many graph algorithms can be viewed as sets of rules that are iteratively applied, with the number of iterations dependent on the size and complexity of the input graph. Existing machine learning architectures often struggle to represent these algorithmic decisions as discrete state transitions. Therefore, we propose a novel framework: GraphFSA (Graph Finite State Automaton). GraphFSA is designed to learn a finite state automaton that runs on each node of a given graph. We test GraphFSA on cellular automata problems, showcasing its abilities in a straightforward algorithmic setting. For a comprehensive empirical evaluation of our framework, we create a diverse range of synthetic problems. As our main application, we then focus on learning more elaborate graph algorithms. Our findings suggest that GraphFSA exhibits strong generalization and extrapolation abilities, presenting an alternative approach to represent these algorithms.
  • One Shot vs. Iterative: Rethinking Pruning Strategies for Model Compression
    Item type: Conference Paper
    Janusz, Mikołaj; Wojnar, Tomasz; Li, Yawei; et al. (2025)
    Frontiers in Artificial Intelligence and Applications ~ ECAI 2025
    Pruning is a core technique for compressing neural networks to improve computational efficiency. This process is typically approached in two ways: one-shot pruning, which involves a single pass of training and pruning, and iterative pruning, where pruning is performed over multiple cycles for potentially finer network refinement. Although iterative pruning has historically seen broader adoption, this preference is often assumed rather than rigorously tested. Our study presents one of the first systematic and comprehensive comparisons of these methods, providing rigorous definitions, benchmarking both across structured and unstructured settings, and applying different pruning criteria and modalities. We find that each method has specific advantages: one-shot pruning proves more effective at lower pruning ratios, while iterative pruning performs better at higher ratios. Building on these findings, we advocate for patience-based pruning and introduce a hybrid approach that can outperform traditional methods in certain scenarios, providing valuable insights for practitioners selecting a pruning strategy tailored to their goals and constraints. Source code is available at https://github.com/janumiko/pruning-benchmark.
Publications 1 - 9 of 9