A group of researchers backed by the French government have released what is thought to be the largest AI training dataset composed entirely of text that is in the public domain. "There's no fundamental reason why someone couldn't train an LLM fairly," says Ed Newton-Rex, CEO of Fairly Trained. He founded the nonprofit in January 2024 after quitting his executive role at image generation startup Stability AI because he disagreed with its policy of scraping content without permission. Fairly Trained offers a certification to companies willing to prove that they've trained their AI models on data that they either own, have licensed, or is in the public domain. When the nonprofit launched, some critics pointed out that it hadn't yet identified a large language model that met those requirements.

At an AI research lab on the edges of Abu Dhabi last year, an international team of 25 computer scientists were putting the finishing touches on a deep learning algorithm before sending it to be trained on 4,000 powerful computer chips. The AI system, which cost several million dollars to train, was funded by an arm of the Abu Dhabi government called the Advanced Technology Research Council (ATRC). Despite the government's substantial investment, ATRC director Faisal Al Bannai decided to release the finished model online for free. If it was as good as the team believed, the boost to the United Arab Emirates' reputation would be all the return the government needed on its investment, he reasoned. When the AI, named Falcon after the UAE's national bird, was publicly released last September, it became a sensation. By some measures it was the best open-source large language model (LLM) available in the world at that point, outperforming top offerings from Meta and Google.

Microsoft has appointed the co-founder of the British artificial intelligence lab DeepMind as the head of a new AI division. Mustafa Suleyman, 39, co-founded DeepMind with Demis Hassabis and Shane Legg in 2010 and the company went on to be bought by Google for 400m in 2014. It now forms the core of Google's AI efforts after merging with another unit to become Google DeepMind in 2023. The chief executive of Microsoft, Satya Nadella, announced in a blogpost that the British AI pioneer, who left DeepMind in 2019, will be chief executive of a new organisation called Microsoft AI focusing on the US company's consumer products and research. Several employees at Sulyeman's Inflection AI startup will join the division.

Microsoft has named Mustafa Suleyman head of its consumer artificial intelligence business, hiring most of the staff from his Inflection AI startup as the software giant seeks to fend off Alphabet's Google in the fiercely contested market for AI products. Suleyman, who co-founded Google's DeepMind, will report to CEO Satya Nadella and oversee a range of projects, such as integrating an AI Copilot into Windows and adding conversational elements to the Bing search engine. His hiring will put Microsoft's consumer AI work under one leader for the first time. Inflection has been a rival of Microsoft's key AI partner OpenAI. The company is shifting to selling AI software to businesses but will continue operating its Pi consumer chatbot business for now.

At the screenings of Ryan Gosling's new movie "The Fall Guy" and Sydney Sweeney's "Immaculate" -- headlining events of the first week of this year's South by Southwest (SXSW) conference -- a reel of speakers touting the merits of AI including Peter Deng, head of ChatGPT at OpenAI, was loudly booed by audiences. In other halls of the annual Austin confab, which attracts over 300,000 people each year, the tone was one of heady optimism. Companies were spinning a positive narrative -- pushing the idea that AI wouldn't destroy jobs, but rather represented a once-in-a-generation opportunity to modernize a multitude of industries. International Business Machines Chief Human Resources Officer Nickle LaMoreaux told a SXSW panel that AI will instead alleviate labor shortages that are set to worsen as birth rates decline across developed economies. And while only a tiny percentage of jobs can be completely automated, the vast majority won't disappear -- but will change dramatically.

This paper introduces CLIPSwarm, a new algorithm designed to automate the modeling of swarm drone formations based on natural language. The algorithm begins by enriching a provided word, to compose a text prompt that serves as input to an iterative approach to find the formation that best matches the provided word. The algorithm iteratively refines formations of robots to align with the textual description, employing different steps for "exploration" and "exploitation". Our framework is currently evaluated on simple formation targets, limited to contour shapes. A formation is visually represented through alpha-shape contours and the most representative color is automatically found for the input word. To measure the similarity between the description and the visual representation of the formation, we use CLIP [1], encoding text and images into vectors and assessing their similarity. Subsequently, the algorithm rearranges the formation to visually represent the word more effectively, within the given constraints of available drones. Control actions are then assigned to the drones, ensuring robotic behavior and collision-free movement. Experimental results demonstrate the system's efficacy in accurately modeling robot formations from natural language descriptions. The algorithm's versatility is showcased through the execution of drone shows in photorealistic simulation with varying shapes. We refer the reader to the supplementary video for a visual reference of the results.

Despite differing from the human language processing mechanism in implementation and algorithms, current language models demonstrate remarkable human-like or surpassing language capabilities. Should computational language models be employed in studying the brain, and if so, when and how? To delve into this topic, this paper reviews efforts in using computational models for brain research, highlighting emerging trends. To ensure a fair comparison, the paper evaluates various computational models using consistent metrics on the same dataset. Our analysis reveals that no single model outperforms others on all datasets, underscoring the need for rich testing datasets and rigid experimental control to draw robust conclusions in studies involving computational models.

Large language models (LLMs) have reached human-like proficiency in generating diverse textual content, underscoring the necessity for effective fake text detection to avoid potential risks such as fake news in social media. Previous research has mostly tested single models on in-distribution datasets, limiting our understanding of how these models perform on different types of data for LLM-generated text detection task. We researched this by testing five specialized transformer-based models on both in-distribution and out-of-distribution datasets to better assess their performance and generalizability. Our results revealed that single transformer-based classifiers achieved decent performance on in-distribution dataset but limited generalization ability on out-of-distribution dataset. To improve it, we combined the individual classifiers models using adaptive ensemble algorithms, which improved the average accuracy significantly from 91.8% to 99.2% on an in-distribution test set and from 62.9% to 72.5% on an out-of-distribution test set. The results indicate the effectiveness, good generalization ability, and great potential of adaptive ensemble algorithms in LLM-generated text detection.

The rapid expansion of the open-source language model landscape presents an opportunity to merge the competencies of these model checkpoints by combining their parameters. Advances in transfer learning, the process of fine-tuning pretrained models for specific tasks, has resulted in the development of vast amounts of task-specific models, typically specialized in individual tasks and unable to utilize each other's strengths. Model merging facilitates the creation of multitask models without the need for additional training, offering a promising avenue for enhancing model performance and versatility. By preserving the intrinsic capabilities of the original models, model merging addresses complex challenges in AI - including the difficulties of catastrophic forgetting and multitask learning. To support this expanding area of research, we introduce MergeKit, a comprehensive, open-source library designed to facilitate the application of model merging strategies. MergeKit offers an extensible framework to efficiently merge models on any hardware, providing utility to researchers and practitioners. To date, thousands of models have been merged by the open-source community, leading to the creation of some of the worlds most powerful open-source model checkpoints, as assessed by the Open LLM Leaderboard. The library is accessible at https://github.com/arcee-ai/MergeKit.

Despite the growing capabilities of large language models, there exists concerns about the biases they develop. In this paper, we propose a novel, automated mechanism for debiasing through specified dataset augmentation in the lens of bias producers and in the context of 'restricted industries' with limited data. We additionally create two new additional metrics, the mb-index and db-index, to quantify bias, considering the idea that bias occurs due to both intrinsic model architecture and dataset.