Recipe Recommendation With Hierarchical Graph Attention Network

Recipe recommendation systems play an important role in helping people find recipes that are of their interest and fit their eating habits. Unlike what has been developed for recommending recipes using content-based or collaborative filtering approaches, the relational information among users, recipes, and food items is less explored. In this paper, we leverage the relational information into recipe recommendation and propose a graph learning approach to solve it. In particular, we propose HGAT, a novel hierarchical graph attention network for recipe recommendation. The proposed model can capture user history behavior, recipe content, and relational information through several neural network modules, including type-specific transformation, node-level attention, and relation-level attention. We further introduce a ranking-based objective function to optimize the model. Thorough experiments demonstrate that HGAT outperforms numerous baseline methods.

High content Image Analysis to study phenotypic heterogeneity in endothelial cell monolayers

Endothelial cells (EC) are heterogeneous across and within tissues, reflecting distinct, specialised functions. EC heterogeneity has been proposed to underpin EC plasticity independently from vessel microenvironments. However, heterogeneity driven by contact-dependent or short-range cell-cell crosstalk cannot be evaluated with single cell transcriptomic approaches as spatial and contextual information is lost. Nonetheless, quantification of EC heterogeneity and understanding of its molecular drivers is key to developing novel therapeutics for cancer, cardiovascular diseases and for revascularisation in regenerative medicine. Here, we developed an EC profiling tool (ECPT) to examine individual cells within intact monolayers. We used ECPT to characterise different phenotypes in arterial, venous and microvascular EC populations. In line with other studies, we measured heterogeneity in terms of cell cycle, proliferation, and junction organisation. ECPT uncovered a previously under-appreciated single-cell heterogeneity in NOTCH activation. We correlated cell proliferation with different NOTCH activation states at the single cell and population levels. The positional and relational information extracted with our novel approach is key to elucidating the molecular mechanisms underpinning EC heterogeneity.

Relation extraction in structured and unstructured data: a comparative investigation on smartphone titles in the e-commerce domain

As large amounts of unstructured data are generated on a regular basis, expressing or storing knowledge in a way that is useful remains a challenge. In this context, Relation Extraction (RE) is the task of automatically identifying relationships in unstructured textual data. Thus, we investigated the relation extraction on unstructured e-commerce data from the smartphone domain, using a BERT model fine-tuned for this task. We conducted two experiments to acknowledge how much relational information it is possible to extract from product sheets (structured data) and product titles (unstructured data), and a third experiment to compare both. Analysis shows that extracting relations within a title can retrieve correct relations that are not evident on the related sheet.

A multistage retrieval account of associative recognition ROC curves

Despite its name, associative recognition is a paradigm thought to rely on memory recall. However, it remains unclear how associative information may be represented and retrieved from memory and what its relationship to other information, such as item memory, is. Here, we propose a computational model of associative recognition, where relational information is accessed in a generic, multistage retrieval process. The model explains the relative difficulty of associative recognition compared with item recognition, the difference in experimental outcomes when different types of lures are used, as well as the conditions leading to the emergence of associative ROC curves with different shapes.

The Vacuum-Phonon (Sonon) Reinterpretation of QED

This paper provides a reference copy of one particular and highly informal comment in a multiweek Academia.edu discussion of the paper Randomness in Relational Quantum Mechanics by Gary Gordon. The other main participants in this particular thread of the discussion were Doug Marman, Conrad Dale Johnson, Ruth Kastner, and the author. In this comment, the author argues that the only self-consistent approach to reconciling Feynman path integrals with Maxwell’s experimentally well-proven theory of electromagnetic wave pressure is introducing a new spin-0 particle, the vacuum or space phonon (sonon), that conveys linear momentum. The path histories of QED become the always-expanding structure of the sonon field, which, like a bubble, becomes increasingly unstable as it expands. The collection of all sonon fields around well-defined bundles of conserved quantum properties creates xyz space by defining the complete set of relational information for those entities. Spacetime in the sonon model is granular, multi-scale, and entirely mass-energy dependent. Implications of the sonon model are discussed, including the need for a drastic update to general relativity to take the multi-scale granularity of spacetime directly into account, rather than explaining it obliquely via models such as dark matter, dark energy, or MOND.

Medial temporal and prefrontal cortices encode structural task representations at different levels of abstraction

Memory generalisations may be underpinned by either encoding- or retrieval-based mechanisms. We used a transitive inference task to investigate whether these generalisation mechanisms are influenced by progressive vs randomly interleaved training, and overnight consolidation. On consecutive days, participants learnt pairwise discriminations from two transitive hierarchies before being tested during fMRI. Inference performance was consistently better following progressive training, and for pairs further apart in the transitive hierarchy. BOLD pattern similarity correlated with hierarchical distances in the medial temporal lobe (MTL) and medial prefrontal cortex (MPFC). These results are consistent with the use of representations that directly encode structural relationships between different task features. Furthermore, BOLD patterns in MPFC were similar across the two independent hierarchies. We conclude that humans preferentially employ encoding-based mechanisms to store map-like relational codes that can be used for memory generalisation. While both MTL and MPFC support these representations, the MPFC encodes more abstract relational information.

How development sculpts hippocampal circuits and function

In mammals, the selective transformation of transient experience into stored memory occurs in the hippocampus, which develops representations of specific events in the context in which they occur. In this review, we focus on the development of hippocampal circuits and the self-organized dynamics embedded within them since the latter critically support the role of the hippocampus in learning and memory. We first discuss evidence that adult hippocampal cells and circuits are sculpted by development as early as during embryonic neurogenesis. We argue that these primary developmental programs provide a scaffold onto which later experience of the external world can be grafted. Next, we review the different sequences in the development of hippocampal cells and circuits at anatomical and functional levels. We cover a period extending from neurogenesis and migration to the appearance of phenotypic diversity within hippocampal cells, and their wiring into functional networks. We describe the progressive emergence of network dynamics in the hippocampus, from sensorimotor-driven early sharp waves to sequences of place cells tracking relational information. We outline the critical turn points and discontinuities in that developmental journey, and close by formulating open questions. We propose that rewinding the process of hippocampal development helps understand the main organization principles of memory circuits.

Examining Automatic Stereotyping From a Propositional Perspective: Is Automatic Stereotyping Sensitive to Relational and Validity Information?

Research on automatic stereotyping is dominated by the idea that automatic stereotyping reflects the activation of (group–trait) associations. In two preregistered experiments (total N = 391), we tested predictions derived from an alternative perspective that suggests that automatic stereotyping is the result of the activation of propositional representations that, unlike associations, can encode relational information and have truth values. Experiment 1 found that automatic stereotyping is sensitive to the validity of information about pairs of traits and groups. Experiment 2 showed that automatic stereotyping is sensitive to the specific relations (e.g., whether a particular group is more or less friendly than a reference person) between pairs of traits and groups. Interestingly, both experiments found a weaker influence of validity/relational information on automatic stereotyping than on non-automatic stereotyping. We discuss the implications of these findings for research on automatic stereotyping.

Examining automatic stereotyping from a propositional perspective: Is automatic stereo-typing sensitive to relational and validity information?

Research on automatic stereotyping is dominated by the idea that automatic stereotyping reflects the activation of (group-trait) associations. In two preregistered experiments (total N=391) we tested predictions derived from an alternative perspective that suggests that automatic stereotyping is the result of the activation of propositional representations that, unlike associations, can encode relational information and have truth values. Experiment 1 found that automatic stereotyping is sensitive to the validity of information about pairs of traits and groups. Experiment 2 showed that automatic stereotyping is sensitive to the specific relations (e.g., whether a particular group is more or less friendly than a reference person) between pairs of traits and groups. Interestingly, both experiments found a weaker influence of validity/relational information on automatic stereotyping than on non-automatic stereotyping. We discuss the implications of these findings for research on automatic stereotyping.