Search AI Products and News

Joelle Pineau, Meta's Vice President of AI Research, announced on Tuesday via Facebook that she plans to leave the company in May. A highly respected leader within Meta's AI research lab (FAIR) for over two years, she oversaw the company's innovation and development in AI. Pineau's departure comes at a crucial time for Meta, as the company plans to invest up to $65 billion in AI by 2025.

Joelle Pineau, VP of Meta AI Research, announced on Tuesday via a Facebook post that she will be leaving her position at Meta in May. Pineau has served as the head of Meta's Fundamental AI Research (FAIR) lab for the past two years, leading the lab's cutting-edge research in artificial intelligence. FAIR is a core internal research team at Meta, led by renowned scientist Yann LeCun. Pineau's departure comes as Meta...

In the field of deep learning, normalization layers are considered an indispensable component of modern neural networks. Recently, research led by Meta FAIR research scientist Zhuang Liu, titled "Transformer without Normalization Layers", has garnered significant attention. This research not only introduces a new technique called Dynamic Tanh (DyT), but also demonstrates the effectiveness of Transformer architectures even without traditional normalization layers.

Recently, a study by the BBC has revealed that leading AI assistants often produce misleading and inaccurate content when answering questions related to news and current affairs. The research indicates that over half of the responses generated by four mainstream AI tools—ChatGPT, Gemini, and Perplexity—are deemed to have 'significant issues'. Note: The image is AI-generated, provided by the image licensing service Midjourney, as requested by researchers for these four generative AI tools.

Xiaohongshu blogger 'kilingvirgin' uses AI technology to transform the classic anime 'Sailor Moon' into a dark fairy tale style, creating a unique visual effect and profound artistic atmosphere that quickly attracted a lot of attention. The AI-generated images cleverly blend elements of the original work with the dark style of the Brothers Grimm through blurry textures, dark tones, and mysterious scenes. This strong visual contrast and innovative adaptation have received widespread recognition, evoking nostalgia for classics and dark fairy tale aesthetics. The creative wave originating from TikTok has spread rapidly.

In the field of natural language processing, large language models perform exceptionally well on complex tasks, but model evaluation heavily relies on expensive and time-consuming human-annotated data. As models advance, the utility of existing data declines, necessitating the continuous collection of new data for scalable and sustainable evaluation. The Meta FAIR research team has proposed the 'Self-Taught Evaluator' to address this issue. This innovative approach trains on synthetic data, eliminating the need for human annotation by generating contrasting synthetic preferences.

The Recording Industry Association of America (RIAA) has filed a lawsuit against music generation startups Udio and Suno, accusing them of using copyrighted music to train AI models. Suno admitted to using copyrighted songs for training and claims that this constitutes fair use. However, the RIAA argues that these actions do not align with the principles of fair use and threaten artists' livelihoods. The ruling in this case could set an important precedent for the AI music training field, affecting the future of related companies and industry regulations.

An AI-rewritten and performed version of 'Fairy Tale Town' has gone viral on Douyin, garnering over 730,000 likes. The AI version merges reality and fantasy, attracting considerable attention with its absurd lyrics and remarkable vocal techniques, provoking public reflection on AI creativity. The lyrics feature scenes like 'Winnie the Pooh picking spicy strips' and 'SpongeBob playing mahjong,' contrasting sharply with the original version and showcasing AI's unique potential in music creation. This viral success not only highlights the infinite possibilities of AI in the creative field but also reflects the public's broad interest in AI artistic creation.

The application of artificial intelligence in job recruitment aims to improve objectivity and efficiency, but studies reveal its potential to exacerbate bias. AI learns from data, and biases in datasets can lead to unfair algorithmic outcomes. The educational and professional gaps between HR and AI developers hinder effective collaboration, impacting the ability to identify and reduce biases. To address this issue, it is recommended to implement structured training programs to enhance HR professionals' knowledge of AI; promote interdisciplinary teamwork to ensure datasets are diverse and represent different groups; and establish guidelines and ethics.

Kingsoft Office announced the WPS AI 2.0 version at the 2024 World Artificial Intelligence Conference and launched its self-developed model for government affairs - the Kingsoft Government Office Model 1.0. The upgrade of WPS AI 2.0 aims to enhance office productivity for users and clients through innovative applications in the personal, enterprise, and government versions. WPS AI Office Assistant has added four AI assistants, including AI Writing Assistant, AI Reading Assistant, AI Data Assista

The counter-cyclical boom in the AI sector has raised concerns among many investors, with early-stage funding globally declining by 40% in 2023, while funding in the AI sector approaches $50 billion.

The nonprofit organization Fairly Trained plans to introduce a certification label to ensure that content generated by AI models does not infringe on copyright. Fairly Trained has approved and issued a label called 'Licensed Model.' The use of copyright-protected data for model training has been a contentious issue, particularly concerning generative AI and its training on copyrighted data.

Ed Newton-Rex, the audio head of Stability AI, resigned due to his disagreement with the company's stance that training generative AI models on copyrighted works falls under 'fair use.' This move sparked widespread discussion about the concept of 'fair use' and the extensive use of content by AI models without creator permission. Newton-Rex opposes the entire industry group adopting the same approach, arguing that copyright issues extend beyond Stability AI.

The UK government is using deep learning algorithms across multiple departments, encountering issues with biased training data. The Home Office's use of AI to read passports at airports has shown significant racial bias. To avoid bias, more transparent algorithms and open data access for expert verification of the system's fairness are needed. While the UK government has taken measures, there is still a need for stronger legal pressure to ensure organizations using AI comply with regulations. We need to ensure fairness and justice in AI technologies through legal, technical, and educational means.