If robust general-purpose reasoning abilities have emerged in LLMs, this bolsters the claim that such systems are an important step on the way to trustworthy general intelligence.

While large language models (LLMs) are not explicitly trained to reason, they have exhibited “emergent” behaviors that sometimes look like reasoning.

The word “reasoning” is an umbrella term that includes abilities for deduction, induction, abduction, analogy, common sense, and other “rational” or systematic methods for solving problems. Reasoning is often a process that involves composing multiple steps of inference.

LLMs are substantially better at solving problems that involve terms or concepts that appear more frequently in their training data, leading to the hypothesis that LLMs do not perform robust abstract reasoning to solve problems, but instead solve problems (at least in part) by identifying patterns in their training data that match, or are similar to, or are otherwise related to the text of the prompts they are given.

[Memorization and reasoning are] not a dichotomy, but rather they can co-exist in a continuum.

Since RDF/XMLlacks a way to express inverse properties

sparql-unlimited

errors in the published data may be caused by either negligence whenfilling out Journal forms or by deliberate obfuscation of sensitive informationin order to reduce a contract’s traceability

LinkedPipes-ETL

The selected datase

sparql-to-graphviz

We extracted the code list standardizing categories of procured ser-vices as defined in the EU directive 2004/18/EC (EU2004). This code listlinks CPV to the Central Product Classification (CPC).8We also extractedseveral code lists enumerating the types of contract notices.

We represent contract notices as instances of the subclasses ofpproc:Noticefrom PPROC, since PCO does not include the concept of a contract notice.PCO treats notices as mere artefacts of the document-based communicationin public procuremen

Infor-mation resource is“a resource which has the property that all of its essentialcharacteristics can be conveyed in a message”

Contract objects expressed viapc:mainObjectandpc:additionalObjectare qual-ified instead of linking CPV directly. A proxy concept that links a CPVconcept viaskos:closeMatchis created for each contract object to allow qual-ification by concepts from the CPV’s supplementary vocabulary. The proxyconcepts link their qualifier viaskos:related.

Thedata-driven approach“implies that vocabularies should not use conceptual-izations that do not match well to common database schemas in their targetdomains”

Public Contracts Ontology3(PCO) is a lightweight RDF vocabulary for de-scribing data about public contracts.

For example, publicprocurement can be used for purchases of drugs in hospitals, cater for roadrepairs, or arrange supplies of electricity

While traditional recommender systems are mostly un-aware of the semantics of the features they use, linked data features do nothave to be treated like black boxes, since their expected interpretations canbe looked up in the corresponding RDF vocabularies that define the features

The principleofAnyone can say anything about anything(AAA) assumes that the openworld of linked data provides no guarantees that the assertions published aslinked data are consistent or uncontradictory

Open world assumption(OWA) supposes that“the truth of astatement is independent of whether it is known. In other words, not know-ing whether a statement is explicitly true does not imply that the statement isfalse

he key challengesof matchmaking by these technologies involve feature selection or featureconstruction, ranking by feature weights, and combination functions for ag-gregating similarity scores of matches

In questo senso, essi non possono essere propriamente considerati Open Data, per i quali tali restrizionitemporali non si applicano.

prevede la possibilità di richiedere esplicitamente dati pubblici non ancora disponibili