China’s cyber-space regulator has unveiled draft measures that would make companies responsible for the data used to train generative artificial intelligence (AI) models, such as Midjourney and ChatGPT.
The tool — dubbed ‘AI-Descartes’ by the researchers — aims to speed up scientific discovery by leveraging symbolic regression, which finds equations to fit data.
Given basic operators, such as addition, multiplication, and division, the systems can generate hundreds to millions of candidate equations, searching for the ones that most accurately describe the relationships in the data.
Using this technique, the AI tool has been able to re-discover, by itself, fundamental equations, including Kepler’s third law of planetary motion; Einstein’s relativistic time-dilation law, and Langmuir’s equation of gas adsorption.
Any activity that requires comprehension and production in one or more modalities is considered a multimodal task; these activities can be extremely varied and lengthy. It is challenging to scale previous multimodal systems because they rely heavily on gathering a large supervised training set and developing task-specific architecture, which must be repeated for every new task. In contrast, present multimodal models have not mastered people’s ability to learn new tasks in context, meaning that they can do so with minimal demonstrations or instructions. Generative pretrained language models have recently shown impressive skills in learning from context.
New research by researchers from Beijing Academy of Artificial Intelligence, Tsinghua University, and Peking University introduces Emu2, a 37-billion-parameter model, trained and evaluated on several multimodal tasks. Their findings show that when scaled up, a multimodal generative pretrained model can learn similarly in context and generalize well to new multimodal tasks. The objective of the predict-the-next-multimodal-element (textual tokens or visual embeddings) is the only one used during Emu2’s training. This unified generative pretraining technique trains models by utilizing large-scale multimodal sequences, such as text, image-text pairs, and interleaved image-text video.
The Emu2 model is generative and multimodal; it learns in a multimodal setting to predict the next element. Visual Encoder, Multimodal Modeling, and Visual Decoder are the three main parts of Emu2’s design. To prepare for autoregressive multimodal modeling, the Visual Encoder tokenizes all input images into continuous embeddings, subsequently interleaved with text tokens. The Visual Decoder turns the regressed visual embeddings into a movie or image.
Surveys from business leaders show that they are now doing mass layoffs due to adoption of AI.
A recent survey of 750 business leaders reveals a growing trend: AI replacing jobs. In 2023, 37% of these leaders acknowledged AI-induced layoffs, with 44% expecting more in 2024. Major companies like Paytm and Google are at the forefront, integrating AI to enhance efficiency but at the cost of human jobs. Paytm’s recent layoffs post AI implementation and Google’s potential restructuring of its ad sales unit due to AI advancements highlight this shift.\
\
#ai #job #layoffs \
\
About Channel: \
\
WION The World is One News examines global issues with in-depth analysis. We provide much more than the news of the day. Our aim is to empower people to explore their world. With our Global headquarters in New Delhi, we bring you news on the hour, by the hour. We deliver information that is not biased. We are journalists who are neutral to the core and non-partisan when it comes to world politics. People are tired of biased reportage and we stand for a globalized united world. So for us, the World is truly One.\
\
Please keep discussions on this channel clean and respectful and refrain from using racist or sexist slurs and personal insults.\
\
That Sports Illustrated debacle is only the start of a massive artificial intelligence disruption of the journalism industry.
A tool made by grad students is the ultimate “geoguesser,” using its knowledge of Google Street View to track down any photo’s location.
A new ISO standard aims to provide an overarching framework for the responsible development of AI.
The International Organization for Standardization (ISO) and the International Electrotechnical Commission (IEC) have approved a new international standard, ISO/IEC 42001. This standard is designed to help organizations develop and use AI systems responsibly.
ISO/IEC 42,001 is the world’s first standard for AI management systems and is intended to provide useful guidance in a rapidly evolving technology area. It addresses various challenges posed by AI, such as ethical considerations, transparency and continuous learning. For organizations, the standard is intended to provide a structured way to balance the risks and opportunities associated with AI.
The release of Transformers has marked a significant advancement in the field of Artificial Intelligence (AI) and neural network topologies. Understanding the workings of these complex neural network architectures requires an understanding of transformers. What distinguishes transformers from conventional architectures is the concept of self-attention, which describes a transformer model’s capacity to focus on distinct segments of the input sequence during prediction. Self-attention greatly enhances the performance of transformers in real-world applications, including computer vision and Natural Language Processing (NLP).
In a recent study, researchers have provided a mathematical model that can be used to perceive Transformers as particle systems in interaction. The mathematical framework offers a methodical way to analyze Transformers’ internal operations. In an interacting particle system, the behavior of the individual particles influences that of the other parts, resulting in a complex network of interconnected systems.
The study explores the finding that Transformers can be thought of as flow maps on the space of probability measures. In this sense, transformers generate a mean-field interacting particle system in which every particle, called a token, follows the vector field flow defined by the empirical measure of all particles. The continuity equation governs the evolution of the empirical measure, and the long-term behavior of this system, which is typified by particle clustering, becomes an object of study.
A student project has revealed yet another power of artificial intelligence — it can be extremely good at geolocating where photos are taken.
Three Stanford graduate students built an AI tool to find a location by looking at pictures. Civil rights advocates warn more advanced versions will further erode online privacy.
Continue reading “Stanford Students Develop AI That Can Pinpoint Your Photo Locations” »
These compounds can kill methicillin-resistant Staphylococcus aureus (MRSA), a bacterium that causes deadly infections.
Using artificial intelligence, MIT researchers discovered a class of compounds that can kill methicillin-resistant Staphylococcus aureus (MRSA), a drug-resistant bacterium that causes more than 10,000 deaths in the U.S. each year.
