Curated content for content, computing, and digital experience professionsals

Tag: machine learning (Page 1 of 4)

Lucidworks announces advanced linguistics package

Lucidworks announced the Advanced Linguistics Package for Lucidworks Fusion to power personalized search for users in Asian, European, and Middle Eastern markets. Lucidworks now embeds text analytics from Basis Technology, provider of AI for natural language processing. According to the companies, building, testing, and maintaining the many algorithms and models required to properly support each language is challenging and expensive. Asian, Middle Eastern, and certain European languages require additional processes to handle unique linguistic phenomena, such as lack of whitespace, compound words, and multiple forms of the same word. The combination of Basis with the AI-powered search platform of Lucidworks Fusion is expected to provide accuracy and performance enhancements in information retrieval for the digital experience. Lucidworks’ Advanced Linguistics Package provides language processing in more than 30 languages and advanced entity extraction in 21 languages. By accurately analyzing the text, in the language it was written, Rosette helps the Lucidworks Fusion platform deliver the right answers to every user, regardless of where they work or what language they use.

https://lucidworks.comhttps://www.basistech.com

Luminoso announces enhancements to open data semantic network

Luminoso, who turn unstructured text data into business-critical insights, announced the newest features of ConceptNet, an open data semantic network whose development is led by Luminoso Chief Science Officer Robyn Speer. ConceptNet originated from MIT Media Lab’s Open Mind Common Sense project more than two decades ago, and the semantic network is now used in AI applications around the world. ConceptNet is cited in more than 700 AI papers in Google Scholar, and its API is queried over 500,000 times per day from more than 1,000 unique IPs. Luminoso has incorporated ConceptNet into its proprietary natural language understanding technology, QuickLearn 2.0. ConceptNet 5.8 features:

Continuous deployment: ConceptNet is now set up with continuous integration using Jenkins and deployment using AWS Terraform, which will make it faster to deploy new versions of the semantic network and easier for others to set up mirrors of the API.

Additional curation of crowd-sourced data: ConceptNet’s developers have filtered entries from Wiktionary that were introducing hateful terminology to ConceptNet without its context. This is part of their ongoing effort to prevent human biases and prejudices from being built into language models. ConceptNet 5.8 has also updated its Wiktionary parser so that it can handle updated versions of the French and German-language Wiktionary projects.

HTTPS support: Developers can now reach ConceptNet’s website and API over HTTPS, improving data transfer security for applications using ConceptNet.

http://blog.conceptnet.io/posts/2020/conceptnet-58/, https://luminoso.com/how-it-works

Gilbane Advisor 5-5-20 — no proof, medium hard, build it, pod-mail?

A radical solution to scale AI technology

Skip the proof of concept? This isn’t, or shouldn’t, be radical. It’s often a good idea for large scale projects, and not just for AI, or other digital experience or content technology initiatives.

Illustration: Israel G. Vargas
scaling AI
 
The example in this article is a customer experience chatbot for Nordea. Read More

How Medium became the best and worst place for coronavirus news

Medium’s pivots over the years created confusion about what they are and who they are for. The editorial challenges inherent in being both a platform and a publisher have only increased over time. Zoe Schiffer’s topical case study illustrates how difficult this balance is. Read More

It’s time to build

If you haven’t read this recent post by Marc Andreessen you should. Though prompted by frustration over our collective response to the current coronavirus pandemic, his prescription for preventing such future failures addresses a broader set of societal problems. Some he mentions; others are implicit, or follow, such as the focus on rent-seeking of wall street, VCs, and, well, too many of us. Read More

The New York Times’ morning email newsletter is getting an official “host and anchor”

Joshua Benton asks “Can any of the lessons of The Daily’s success be carried over into your inbox?” and attempts an answer, or rather asks the right questions. The new “The Morning” launched this week, and as someone who curates a newsletter I’ll be paying attention. But a podcast and an email newsletter are very different animals. Read More

Also…

The Gilbane Advisor curates content for content technology, computing, and digital experience professionals. We focus on strategic technologies. We publish more or less twice a month except for August and December.
We do not sell or share personal data.

Subscribe | Feed | View online | Privacy policy | Editorial policy

Bluebill Advisors Inc logo

 

Gilbane Advisor 4-3-20 — WFH, content technology, AGI not

Content and MarTech vendor subway maps

In 2008 Tony Byrne came up with the idea of a “subway map” as a useful and fun way to illustrate the content technology vendor landscape. He has updated the map every year to incorporate the shifting landscape, sprawling growth, adjacent technologies, and of course the renaming and repositioning by vendors and market analysts.  

content technology vendor subway map

In this article, he shares all 12 subway maps and his thoughts on the changes each year. History is always relevant. A good read. Read More

Scroll, Firefox and ad-free news

Though their impact may be small, at least to start, the business model is interesting. Read More

Scroll and Firefox no-ad news

RealWorld framework comparison 

Handy up-to-date info for front-end-developers. Comparing performance, size, and lines of code implementing Conduit. Read More

RealWorld frameworks

The end of Starsky Robotics

This is a cautionary tale of what can happen when an enthusiastic founder and hungry investors crank each other up without guarding against mutually assured destructive confirmation bias, and don’t do enough serious due diligence. This scenario is unfortunately common, though often with enough funding/time/expert support a pivot or two can prevent disaster.

In this particular case, the problem was a naive expectation of what machine learning could, or would soon be able to, accomplish. Even the possibility of Artificial General Intelligence (AGI) is controversial among experts in the field. I only share this because Starsky’s founder and CEO Stefan Seltz-Axmacher had the courage to publish it. Kudos to him for sharing what happened, and providing enough detail for a valuable case study for entrepreneurship programs. Read More

Also…

The Gilbane Advisor curates content for content technology, computing, and digital experience professionals. We focus on strategic technologies. We publish more or less twice a month except for August and December. We do not sell or share personal data.

Subscribe | Feed | View online | Privacy policy | Editorial policy

 

 

Gilbane Advisor 3-4-20 — IKEA, T5, AI

IKEA sets a new privacy standard for marketers

Tim Walters reports on an impressive approach by IKEA to earn consumer’s trust, by doing rather than (just) promising. As Tim says, you should really watch the IKEA video description and demo with their delightfully down-to-earth Chief Digital Officer, Barbara Martin Coppola. Read More

Transfer Learning with T5: the Text-To-Text Transfer Transformer

Many of you are familiar with natural language processing (NLP) from the rule-based machine translation in the 80s to today’s more successful machine learning approaches. This post from the Google AI Blog describes a promising new Transfer Learning technique and openly available tools. Slightly technical with a link to the academic paper.

With T5, we propose reframing all NLP tasks into a unified text-to-text-format where the input and output are always text strings, in contrast to BERT-style models that can only output either a class label or a span of the input. Our text-to-text framework allows us to use the same model, loss function, and hyperparameters on any NLP task, including machine translation, document summarization, question answering, and classification tasks (e.g., sentiment analysis). Read More

The new business of AI (and how it’s different from traditional software)

Martin Casado and Matt Bornstein from Andreessen Horowitz wrote a thoughtful piece for AI startups and investors on the differences between the business models of AI companies and software companies. As investors they have a particular interest in the margin potential look at the resources and costs associated with each. My take is that is that they have identified a difference of degree rather than of kind, at least in the case of enterprise software applications, which have similar scaling, “humans in the loop”, interoperability, custom development, and support requirements. Large scale content management systems and “digital experience platforms” are examples. In any case, this is a good read, and all the authors’ recommendations should also be considered by traditional enterprise software companies :).  Read More

Update on technology transformations

McKinsey reports on enterprise’s view and appetite for continued technology transformation. Tldr; it’s hard but showing benefits, and competitiveness demands its continuation. Read More

Business side support IT in top companies

Also…

The Gilbane Advisor curates content for content technology, computing, and digital experience professionals. We focus on strategic technologies. We publish more or less twice a month except for August and December. We do not sell or share personal data.

« Older posts