Goto

Collaborating Authors

 own word


Measuring (a Sufficient) World Model in LLMs: A Variance Decomposition Framework

arXiv.org Artificial Intelligence

Understanding whether large language models (LLMs) possess a world model-a structured understanding of the world that supports generalization beyond surface-level patterns-is central to assessing their reliability, especially in high-stakes applications. We propose a formal framework for evaluating whether an LLM exhibits a sufficiently robust world model, defined as producing consistent outputs across semantically equivalent prompts while distinguishing between prompts that express different intents. We introduce a new evaluation approach to measure this that decomposes model response variability into three components: variability due to user purpose, user articulation, and model instability. An LLM with a strong world model should attribute most of the variability in its responses to changes in foundational purpose rather than superficial changes in articulation. This approach allows us to quantify how much of a model's behavior is semantically grounded rather than driven by model instability or alternative wording. We apply this framework to evaluate LLMs across diverse domains. Our results show how larger models attribute a greater share of output variability to changes in user purpose, indicating a more robust world model. This improvement is not uniform, however: larger models do not consistently outperform smaller ones across all domains, and their advantage in robustness is often modest. These findings highlight the importance of moving beyond accuracy-based benchmarks toward semantic diagnostics that more directly assess the structure and stability of a model's internal understanding of the world.


Can a Chatbot Publish an "Original" Novel?

Slate

This story is part of Future Tense Fiction, a monthly series of short stories from Future Tense and Arizona State University's Center for Science and the Imagination about how technology and science will change our lives. THE COURT: Please be seated. Let's try to keep the temperature down in here. We don't need a repeat of yesterday. It'll just be Mr. Blatz and myself today. Sorry, it's hard to tell with โ€ฆ are you with us? ORWELL: Omni-dimensional Recursively Written Entity for Language Learning present and ready, Your Honor. THE COURT: You can just say ORWELL. Are we ready to proceed? LIU: Your Honor, we'd like to call the Defendant to the stand. Mr. Blatz will handle examination. THE COURT: We have the wiring sorted out? Please refrain from using the monitor on the Defendant's table until you're off the stand.


How to Use Voice AI to Empower Live Agents

#artificialintelligence

Your call center agents are the lifeblood of your contact center. They have a huge responsibility to represent your brand during a customer's time of need and frustration, all while actually solving a customer service issue in real-time. Whatsmore, positive and negative experiences with your call center agents have a direct impact on your bottom line, resulting in either customer retention or customer churn. In fact, 90% of Americans say that their customer service experience directly influences their decision to either continue or stop doing business with a brand. With fierce competition, it's important to make sure your call center agents are empowered and set up for success.


Content rewriting techniques using NLP paraphrasers

#artificialintelligence

Content assists you in achieving your objectives and is regarded as the essential factor in many fields. Content can be many things like a thesis for students, any blog post to engage bloggers, or a marketing post to attract customers for digital businesses. Unique and valuable content is what people want from you. Content that is unique and comprehensive is a core in any type of writing, be it an article, blog, or scholarly document. This will help you build your reputation while also preventing you from being accused of Plagiarism.


Five Artificial Intelligence Insiders in Their Own Words

#artificialintelligence

Second, good design can bring A.I.'s benefits to those who otherwise might be left out. That so much of A.I. is currently directed at the affluent contradicts the notion that good design can serve everyone, regardless of age, condition or economic background. We should take the "A.I. for all" approach, and it should follow a human need. Designers and researchers should work together in the early stages of design to identify those needs, develop A.I. that responds compassionately to human demands, and use design to ensure cost-effective, accessible A.I.-enabled products and services. Third, A.I. should never create emotional dependence.


Impressions of @Scale 2018 conference โ€“ Becoming Human: Artificial Intelligence Magazine

#artificialintelligence

The day started with 3 keynotes. David Patterson received a Turing Award in 2017 for his contribution in design and evaluation of computer architectures. In the talk he speaks about the end of Moore's law and it's impact on the industry. Given that hardware alone is not providing that much speed up anymore David Patterson sees a potential solution in creating Domain-Specific Architectures (DSA) and using Domain-Specific Languages (DSL) as a way to get high performance computation. As an example, he showed matrix multiplication in Python which can be speed up about 63000 times(!) if written in C and takes advantage of memory optimizations and vectorized instructions.


SoftBank's Masayoshi Son in his own words -- all 303,513 of them

The Japan Times

Masayoshi Son has a lot going on these days. The SoftBank Group Corp. founder oversees the largest technology fund in history and is putting money into ride-hailing, e-commerce, digital payments, satellites, semiconductors, agriculture and cancer detection -- just for a start. How to make sense of it all? Son uses earnings briefings to explain his evolving vision. So ahead of SoftBank's first-quarter results on Monday, Bloomberg News analyzed his comments from previous presentations to see how his focus has changed and what may lay ahead.


Diffrential Privacy with AI and the future of health tracking

#artificialintelligence

One of the major reasons people are vary of jumping onto the self-tracking wagon is privacy of their data. Granted, when you're trusting a fitness tracker to record how much you've run and how your heart rate varies, you're trusting a slew of services that are working in the background. How can that balance be achieved? Now with iOS 10, Apple first introduced the concept of differential privacy. And trust Apple to come up with a solution to the whole data gathering / privacy conundrum. While the usage of data driven AI services implies that data collection is necessary, differential privacy brings an entirely new level of identity masking into play.


5 People Who Are Pushing Technology That You Will Want To Work With Now

Forbes - Tech

Here are some incredibly smart people that are pushing boundaries and/or doing some incredibly inventive things in emerging technology and other areas. Whether it's healthcare, news or immersive technologies - these are the people you'll be hearing a lot more from - collaborate with them now and be a bigger part of the future. Burrell was described as "offering new thinking and fresh ideas to strengthen the future of broadcasting" after taking top honors at the National Association of Broadcasters Innovation Pilot Award for "In Your Shoes" (a broadcast and 360 VR series), Recently working with heavy-hitters like Steven Spielberg's virtual reality company ('The VR Company'), Burrell's work with Oculus has been considered for an Interactive Emmy ("Take Back the Mic: The World Cup of Hip Hop"). Recently returning from South Africa where she created a partnership with Ndaba Mandela (grandson to Nelson Mandela and Founder of Africa Rising), to bring the coding, robotics and emerging technology program permanently to the continent. IN HER OWN WORDS: "There's a strong feeling of uncertainty in the world right now, but when we ask ourselves, 'how can I use the the tools of today and the technology of tomorrow to harness something meaningful and good', we are empowered to reshape our experience of the world."


Carrie Fisher in her own words

BBC News

A selection of quotes from Carrie Fisher, the Star Wars actress, novelist and screenwriter, who has died aged 60. "When I got the part of a princess in this goofy little science-fiction film, I thought: it'll be fun to do. I'm 19! Who doesn't want to have fun at 19? "I'll go hang out with a bunch of robots for a few months and then return to my life and try to figure out what I want to do when I grow up. "But then Star Wars, this goofy, little three-month hang-out with robots did something unexpected. "It exploded across the firmament of pop culture, taking all of us along with it. I'm on seven medications, and I take medication three times a day.