Lifeboat Foundation

When people program new deep learning AI models — those that can focus on the right features of data by themselves — the vast majority rely on optimization algorithms, or optimizers, to ensure the models have a high enough rate of accuracy. But one of the most commonly used optimizers — derivative-based optimizers— run into trouble handling real-world applications.

In a new paper, researchers from DeepMind propose a new way: Optimization by PROmpting (OPRO), a method that uses AI large language models (LLM) as optimizers. The unique aspect of this approach is that the optimization task is defined in natural language rather than through formal mathematical definitions.

The researchers write, “Instead of formally defining the optimization problem and deriving the update step with a programmed solver, we describe the optimization problem in natural language, then instruct the LLM to iteratively generate new solutions based on the problem description and the previously found solutions.”

ChatGPT went down on Wednesday morning — and the timing of its outage couldn’t have been more unfortunate. While OpenAI’s world-beating chatbot suffered its second major outage in as many weeks, big tech executives were convening in Washington to plead their case to lawmakers over the future of AI.

Among several notable figures in attendance was Sam Altman, CEO of the AI startup — who probably hoped to put on a better face amidst increased scrutiny over ChatGPT’s falling user traffic for the past several months.

This was yet another notable outage that ChatGPT has suffered in the past several weeks as user traffic falls.

The CEO of Polish drinks company Dictador is an AI-powered humanoid robot who works 7 days a week. The AI boss, named Mika, told Reuters that she doesn’t have weekends and is “always on 24/7.” Mika helps to spot potential clients and selects artists to design the rum producer’s bottles.

The humanoid robot CEO of a Polish drinks company is one busy boss.

Continue reading “The humanoid robot CEO of a drinks company says she doesn’t have weekends and is ‘always on 24/7’” »

Imagine living in a cool, green city flush with parks and threaded with footpaths, bike lanes, and buses, which ferry people to shops, schools, and service centers in a matter of minutes.

That breezy dream is the epitome of urban planning, encapsulated in the idea of the 15-minute city, where all basic needs and services are within a quarter of an hour’s reach, improving public health and lowering vehicle emissions.

Artificial intelligence could help urban planners realize that vision faster, with a new study from researchers at Tsinghua University in China demonstrating how machine learning can generate more efficient spatial layouts than humans can, and in a fraction of the time.

Get up to speed on the rapidly evolving world of AI with our roundup of the week’s developments.

In a move that should surprise no one, tech leaders who gathered at closed-door meetings in Washington, DC, this week to discuss AI regulation with legislators and industry groups agreed on the need for laws governing generative AI technology. But they couldn’t agree on how to approach those regulations.

“The Democratic senator Chuck Schumer, who called the meeting ‘historic,’ said that attendees loosely endorsed the idea of regulations but that there was little consensus on what such rules would look like,” The Guardian reported. “Schumer said he asked everyone in the room — including more than 60… More.

In tasks like customer service, consulting, programming, writing, teaching, etc., language agents can reduce human effort and are a potential first step toward artificial general intelligence (AGI). Recent demonstrations of language agents’ potential, including AutoGPT and BabyAGI, have sparked much attention from researchers, developers, and general audiences.

Even for seasoned developers or researchers, most of these demos or repositories are not conducive to customizing, configuring, and deploying new agents. This restriction results from the fact that these demonstrations are frequently proof-of-concepts that highlight the potential of language agents rather than being more substantial frameworks that can be used to gradually develop and customize language agents.

Furthermore, studies show that the majority of these open-source sources cover only a tiny percentage of the basic language agent abilities, such as job decomposition, long-term memory, web navigation, tool usage, and multi-agent communication. Additionally, most (if not all) of the language agent frameworks currently in use rely exclusively on a brief task description and entirely on the ability of LLMs to plan and act. Due to the high randomness and consistency across different runs, language agents are difficult to modify and tweak, and the user experience is poor.

“Our camera uses a completely new method to achieve high-speed imaging. It has an imaging speed and spatial resolution similar to commercial high-speed cameras but uses off-the-shelf components.”

Scientists from the Institut National De La Recherche Scientifique (INRS) in Canada, in collaboration with Concordia University and Meta Platforms Inc., unveiled a game-changing camera that could revolutionize high-speed imaging.

The diffraction-gated real-time ultrahigh-speed mapping (DRUM) camera, introduced in a recent paper published in Optica, is poised to democratize ultrafast imaging, making it accessible for a wide range of applications.

This device can generate deep-UV light with a very narrow wavelength range that is safe for humans but lethal for germs.

A new device that can generate deep-ultraviolet (UV) light to kill germs without harming humans has been developed by a team of researchers from Osaka University, Japan. The device uses a novel method of combining two visible photons into one deep-UV photon inside a thin waveguide made of aluminum nitride, a material that has nonlinear optical properties.

The research, named “229 nm far-ultraviolet second harmonic generation in a vertical polarity inverted AlN bilayer channel waveguide,” has been published in the journal Applied Physics Express.

Users should continue to evaluate their settings to understand how they are being tracked.

Following a lengthy examination into its data practices, Google has agreed to pay $93 million to resolve claims that it misled customers about how and when their location information was being tracked and stored. The investigation was led by the California Department of Justice.

This is according to a statement released by California Attorney General Rob Bonta’s office on Thursday.

Continue reading “Google fined $93 million for deceiving users about tracking” »