weekend ai reads for 2024-04-12

From drug smugglers to modern-day pirates, maritime crime fighter Dyhia Belhabib introduces Heva: an AI-powered tool that aggregates international criminal records to detect and stop crime that might otherwise get swept away in the tide.

Then, a drone is launched to follow the gunman as he moves through the campus. It gives location information in real-time that can be passed on to police.

To try to better monitor and quantify the problem of overfishing, Global Fishing Watch is now using increasingly sophisticated AI software, and satellite imagery, to globally map the movements of more than 65,000 commercial fishing vessels, both those with - and without - AIS.

The ruling, signed Friday by King County Superior Court Judge Leroy McCullogh and first reported by NBC News, described the technology as novel and said it relies on “opaque methods to represent what the AI model ‘thinks’ should be shown.”

Nobody else took what I was doing seriously. So nobody would want to work with me. I was thought to be a bit eccentric and maybe cranky.

The fraction of workers at businesses that use AI is higher, especially for large businesses and in the Information sector. AI use is higher in large firms but the relationship between AI use and firm size is non-monotonic. In contrast, AI use is higher in young firms although, on an employment-weighted basis, is U-shaped in firm age.

Our goal is a trustworthy set of human-understandable commitments that is backed up by lightweight terms of service and data sharing agreements.

But a cohort of Big Tech vendors and consultancies — called the AI-Enabled ICT Workforce Consortium (ITC) — aims to push back against the notion that AI will lead to job losses, citing the need for re-skilling and upskilling within the information and communication technology (ICT) industry specifically.

The ITC is being led by Cisco with support from Google, Microsoft, IBM (conspicuously), Intel, SAP and Accenture. The ITC’s mandate is to explore AI’s impact on jobs while enabling people to find AI-related training programs and connecting businesses to “skilled and job-ready” workers, a spokesperson told TechCrunch in a briefing.

The software company is offering its network of photographers and artists $120 to submit videos of people engaged in everyday actions such as walking or expressing emotions including joy and anger, according to documents seen by Bloomberg. The goal is to source assets for artificial intelligence training, the company wrote.

Additional automated systems, including one called “Where’s Daddy?” also revealed here for the first time, were used specifically to track the targeted individuals and carry out bombings when they had entered their family’s residences.

Faculty and students expressed divergent opinions, with students being much more optimistic about the benefits of AI than their faculty, who cited concerns about the dangers of delegating writing and critical thinking to AI. Nevertheless, both students and faculty members recognized that AI is an inevitable part of our futures that cannot be ignored. To this point, the data suggest that a viable solution involves opening conversation between faculty and students about the benefits of properly using AI tools and educating students about how to use the tools ethically.

This paper critically examines the transformation of the educational landscape through the integration of generative AI with Open Educational Resources (OER) and Open Educational Practices (OEP). The emergence of AI in content creation has ignited debate regarding its potential to comprehend and generate human language, creating content that is often indistinguishable from that produced by humans. This shift from organic (human-created) to synthetic (AI-created) content presents a new frontier in the educational sphere, particularly in the context of OER and OEP. The paper explores the generative AI’s capabilities in OER and OEP, such as automatic content generation, resource curation, updating existing resources, co-creation and facilitating collaborative learning

Georgia Tech’s supercomputer runs on 20 Nvidia HGX H100 systems, which house 160 of Nvidia’s H100 GPUs, which are high in demand across the tech industry.

Developed by S&P Global’s AI-focused division, Kensho, the benchmarking tool assesses an LLM’s ability to handle tasks such as quantitative reasoning, data extraction from financial documents and demonstrating domain-specific knowledge. The results are then displayed on a leaderboard, providing a transparent view of each model’s capabilities.

What is notable is not the demo specifics; rather, note the role of the LLM: it is not providing information or taking specific actions, but rather serving as a much more accessible natural language interface to surface and collect data that would otherwise take considerably more expertise and time. In other words, it is trustworthy because it is grounded through integration Google is promising with its other enterprise data services.

To address these challenges, we present novel enhancements to the conventional time series Transformer architecture, resulting in our proposed Masked Encoder-based Universal Time Series Forecasting Transformer (Moirai). Trained on our newly introduced Large-scale Open Time Series Archive (LOTSA) featuring over 27B observations across nine domains, Moirai achieves competitive or superior performance as a zero-shot forecaster when compared to full-shot models.

Table 2 and Figure 3 show that our method generally enhances performance across all tasks and LLMs by increasing the ensemble size. Specifically, in arithmetic reasoning tasks, the accuracy gains range from 12% to 24% on the GSM8K and from 6% to 10% on the MATH. In general reasoning tasks, the accuracy gains range from 1% to 4% on the Chess and from 5% to 11% on the MMLU. In code generation task, the accuracy gains range from 4% to 9% on HumanEval. Surprisingly, our method enables a smaller LLM to outperform a larger counterpart by simply scaling up the ensemble size. For instance, the enhanced Llama2-13B model achieves 59% accuracy on the GSM8K dataset, outperforming the Llama2-70B model, which scores 54%.

I repeated the same completion prompt “Intelligence is'“ hundreds of times and used this to peer into the statistical and semantic behavior of chatgpt

You can create a digital replication of your voice using AI. Simply record your voice for a few minutes and then choose any kids book from our library, to create your personalised audiobook spoken in your voice! Perfect for soothing your little ones to sleep.