Google Search Liaison’s Post

#SEO #Google "We know many people, including those in the research community, value seeing previous versions of webpages when available. That’s why beginning today, we're adding links to the Internet Archive's Wayback Machine to our "About this result" panel, to give people quick context and make this helpful information easily accessible through Search." (1) Source & read more (1): https://lnkd.in/dzvhTp5q (Google SearchLiaison on x.com), https://lnkd.in/d9pTvBT8 (Google SearchLiaison on linkedin.com) "Starting today, users everywhere can view archived versions of webpages directly through Google Search, with a simple link to the Internet Archive’s Wayback Machine." (2) "To access this new feature, conduct a search on Google as usual. Next to each search result, you’ll find three dots—clicking on these will bring up the “About this Result” panel. Within this panel, select “More About This Page” to reveal a link to the Wayback Machine page for that website. Through this direct link, you’ll be able to view previous versions of a webpage via the Wayback Machine, offering a snapshot of how it appeared at different points in time." (2) Source & read more (2): https://lnkd.in/dhZjQDs5 (blog.archive.org) #Marketing #DigitalMarketing #MarketingDigitale #MarketingDigital

Google Search adds Internet Archive’s Wayback Machine links to about this page. This enabling searches to view the previous version of a given webpage! How It Works: To access this new feature, conduct a search on Google as usual. Next to each search result, you’ll find three dots - clicking on these will bring up the “About this Result” panel. Within this panel, select “More About This Page” to reveal a link to the Wayback Machine page for that website. #Google #WaybackMachine #InternetArchive

The changing nature of information changes us. Revised Substack post. https://lnkd.in/ex_S65FS

And the web is forgetful... I've been made aware that the full text for "Modularity and Refinement for Inference Systems" (Ganzinger, Ruess, Shankar) is not publicly available anymore. So, instead of emailing individual copies I just reposted this text on ResearchGate https://lnkd.in/d4geKHFb Based on a deconstruction of algorithms into inference, strategy, and indexing we have been developing a theory for combining and refining algorithms. These developments have originally been show-cased for decision procedures only, but they have proven to be much more general in the meantime.

Google just released the FRAMES benchmark dataset for testing Retrieval-Augmented Generation (RAG) systems. But is it the right fit for your business? FRAMES tests a system’s ability to answer 24 complex, multi-hop questions pulling from multiple Wikipedia articles. You can check it out here: https://lnkd.in/gYdEVmNR It’s great for general RAG systems across a wide range of topics like history and science. But for businesses needing a focused RAG system built on their internal knowledge, FRAMES isn’t quite the right fit. That said, the core concept behind it can be applied to any RAG system: creating a dataset of gold answers with relevant links for common prompts you expect and use the dataset to evaluate the performance of your RAG system. How have you approached this with your own RAG projects? Share your thoughts and experience below!

Happy to see a review of my book Semantic Media in the International Journal of Communication. "Iliadis highlights how major tech companies use semantic technologies to control information flow and urges readers to reflect on the broader implications for personal freedom, public discourse, and democratic governance. [...] Overall, the book provides valuable insights into the influence of semantic technologies on media and democracy, underscoring the need to remain vigilant and actively address the challenges posed by technological advancements to protect democratic values." I'll add that if you are at all interested in information organization on platforms, how facts are presented in search results, and think that "algorithms" has become somewhat of a misnomer for the multiple ways that companies orchestrate information for consumers, then this book will be of interest you! https://lnkd.in/enZDN2Un

Good news for materials scientists! 🔬🔬 We have adapted the powerful AI-based image segmentation tools from AI at Meta to work for micrographs. "SAMBA: A Trainable Segmentation Web-App with Smart Labelling" By Ronan Docherty, Isaac Squires, Antonis Vamvakeros, and Sam Cooper, published in Journal of Open Source Software https://lnkd.in/enrsDDQa At each point in the image, the model suggests the region you might be trying to segment! A transformer trained on millions of pictures of cats and sunsets turns out to also be great at distinguished pores from particles! 🐱🐱🌅🌅 The tool is cloud-based, so you can run this powerful model from your laptop. Visit the website to see the great work Ronan did on the user interface: http://sambasegment.com

I was reading this book on social networks by Kleinberg and Easley, "Networks, Crowds, and Markets" and in section 2.3 they explained the concept of Breadth First Search (BFS) in graph theory. A practical example of this could be how google maps use the same approach for determining the shortest route from location A to location B. It is amazing how a network solution can be used in various ways. You can read more about it here: https://lnkd.in/dtJm38ZG

Computer Science / Mathematics folks:  Does anyone have a pointer to the English version of Petersen's 1891 Acta Mathematica paper [1], which is in German? Specifically, I need the discussion of the properties of regular graphs and 1-factor a.k.a. perfect matching (and also other properties for that matter)! Additionally, it is known that this work precedes Berge's 1957 PNAS paper [2] as far as establishing the relation between maximum matching and augmenting paths, a key lemma in Edmonds' Blossom Algorithm, is concerned. I am not sure whether Petersen just "observed" this relation or actually casually "proved" it as a simple consequence of one of their theorems. I very highly appreciate any pointer whatsoever!! PS: Meine Kenntnisse der deutschen Sprache beschränken sich auf diese Aussage, die ich mir von Google übersetzen ließ. [1] - https://lnkd.in/dSav4eea [2] - https://lnkd.in/dzMAKvBk

Do you have your own personal knowledge graph? For millenia (not exaggerating), scholars have kept "commonplace books": notebooks where they would keep notes, quotes, observations, ideas, and much more. Here's a description from Steven Johnson's book Where Good Ideas Come From: “Darwin’s notebooks lie at the tail end of a long and fruitful tradition that peaked in Enlightenment-era Europe, particularly in England: the practice of maintaining a 'commonplace' book. Scholars, amateur scientists, aspiring men of letters—just about anyone with intellectual ambition in the seventeenth and eighteenth centuries was likely to keep a commonplace book. The great minds of the period—Milton, Bacon, Locke—were zealous believers in the memory-enhancing powers of the commonplace book.” Sophisticated indexing schemes were even developed to help keep track and refer back to all the notes they contain. (Old image below is one of John Locke's commonplace book index.) Now, there's no need to be an "aspiring men of letters" to find value in taking notes. These days, our thoughts and notes are spread all over the place: in emails, on our phones, on blogs, here on LinkedIn, etc. But could we get more value from all of this, if we were able to consolidate our data, to search it, and to navigate it through links and topics? I just started using Obsidian (https://obsidian.md/) as a tool to do that. It transforms your notes in your own personal web of information. Requires a bit of discipline to keep track of everything, but seems worth it. It's super cool!