Blog by Ines Almeida

A research paper proposes formal internal audits as a mechanism for technology companies to ensure their artificial intelligence (AI) aligns with ethical priorities before deployment.

A new perspective paper argues for "measuring data" as a critical task to advance responsible AI development. Just as physical objects can be measured, data used to train AI systems should also be quantitatively analyzed to understand its composition.

A study from AI researchers at OpenAI demonstrates how large language models like chatbots can be adapted to reflect specific societal values through a simple "fine-tuning" process.

A new study has shown that transformers can be expressed in a simple logic formalism. This finding challenges the perception that transformers are inscrutable black boxes and suggests avenues for interpreting how they work.

A new study has shown that transformers can be expressed in a simple logic formalism. This finding challenges the perception that transformers are inscrutable black boxes and suggests avenues for interpreting how they work.

A new study provides insights into the complex processes and people behind ML data work.

In a recent paper, researchers Mehtab Khan and Alex Hanna highlight the need for greater scrutiny, transparency, and accountability in how massive datasets for machine learning models are created.

Machine learning models rely heavily on their training datasets, inheriting inherent biases and limitations. This research proposes "datasheets for datasets" increasing transparency and mitigating risks.

A new study reveals troubling bias in AI detectors of machine-generated text against non-native English speakers. The findings raise important questions around AI fairness and underscore the need for more inclusive technologies.

Protecting the Copyright of Large Language Models Using Waterrmarks

Enhancing AI's Compositional Language Skills

Building QA Systems that catch knowledge gaps and avoid misleading users.

Researchers are exploring new techniques to improve AI's ability to grasp diverse sentence structures and indirect meaning.

Making Conversational AI More Natural: Helping Systems Understand Indirect References

In their research paper, AI21 Labs demonstrates that frozen LLMs have untapped potential that can match or exceed fine-tuning approaches, without the downsides.

A new perspective paper argues that we must evaluate LLMs along two dimensions: formal linguistic competence and functional linguistic competence.

LLMs have some serious limitations that constrain their usefulness for real-world applications. To overcome these limitations, AI researchers have proposed a new type of AI system architecture called Modular Reasoning, Knowledge and Language (MRKL).

Researchers are exploring how to combine LLMs with neurosymbolic methods that incorporate logical reasoning and structure.

When designing BloombergGPT, a new large language model optimized for financial data, the researchers at Bloomberg made an important decision around tokenization.

When designing BloombergGPT, a new large language model optimized for financial data, the researchers at Bloomberg made an important decision around tokenization.