DeepSeek reverts to Nvidia for R2 model after Huawei AI chip failsAI Newson August 14, 2025 at 4:04 pm DeepSeek’s plan to train its new AI model, R2, on Huawei’s Ascend chips has failed and forced a retreat to Nvidia while delaying launch. For months, the narrative pushed by Beijing has been one of unstoppable technological progress and a march towards self-sufficiency. However, reality has a habit of biting back. The recent troubles of
The post DeepSeek reverts to Nvidia for R2 model after Huawei AI chip fails appeared first on AI News.
DeepSeek’s plan to train its new AI model, R2, on Huawei’s Ascend chips has failed and forced a retreat to Nvidia while delaying launch. For months, the narrative pushed by Beijing has been one of unstoppable technological progress and a march towards self-sufficiency. However, reality has a habit of biting back. The recent troubles of
The post DeepSeek reverts to Nvidia for R2 model after Huawei AI chip fails appeared first on AI News. Read More
Do Biased Models Have Biased Thoughts?cs.AI updates on arXiv.orgon August 13, 2025 at 4:00 am arXiv:2508.06671v2 Announce Type: replace-cross
Abstract: The impressive performance of language models is undeniable. However, the presence of biases based on gender, race, socio-economic status, physical appearance, and sexual orientation makes the deployment of language models challenging. This paper studies the effect of chain-of-thought prompting, a recent approach that studies the steps followed by the model before it responds, on fairness. More specifically, we ask the following question: $textit{Do biased models have biased thoughts}$? To answer our question, we conduct experiments on $5$ popular large language models using fairness metrics to quantify $11$ different biases in the model’s thoughts and output. Our results show that the bias in the thinking steps is not highly correlated with the output bias (less than $0.6$ correlation with a $p$-value smaller than $0.001$ in most cases). In other words, unlike human beings, the tested models with biased decisions do not always possess biased thoughts.
arXiv:2508.06671v2 Announce Type: replace-cross
Abstract: The impressive performance of language models is undeniable. However, the presence of biases based on gender, race, socio-economic status, physical appearance, and sexual orientation makes the deployment of language models challenging. This paper studies the effect of chain-of-thought prompting, a recent approach that studies the steps followed by the model before it responds, on fairness. More specifically, we ask the following question: $textit{Do biased models have biased thoughts}$? To answer our question, we conduct experiments on $5$ popular large language models using fairness metrics to quantify $11$ different biases in the model’s thoughts and output. Our results show that the bias in the thinking steps is not highly correlated with the output bias (less than $0.6$ correlation with a $p$-value smaller than $0.001$ in most cases). In other words, unlike human beings, the tested models with biased decisions do not always possess biased thoughts. Read More
How We Reduced LLM Costs by 90% with 5 Lines of CodeTowards Data Scienceon August 21, 2025 at 7:08 pm When clean code hides inefficiencies: what we learned from fixing a few lines of code and saving 90% in LLM cost.
The post How We Reduced LLM Costs by 90% with 5 Lines of Code appeared first on Towards Data Science.
When clean code hides inefficiencies: what we learned from fixing a few lines of code and saving 90% in LLM cost.
The post How We Reduced LLM Costs by 90% with 5 Lines of Code appeared first on Towards Data Science. Read More
A ChatGPT-based approach for questions generation in higher educationcs.AI updates on arXiv.orgon July 31, 2025 at 4:00 am arXiv:2507.21174v2 Announce Type: replace-cross
Abstract: Large language models have been widely applied in many aspects of real life, bringing significant efficiency to businesses and offering distinctive user experiences. In this paper, we focus on exploring the application of ChatGPT, a chatbot based on a large language model, to support higher educator in generating quiz questions and assessing learners. Specifically, we explore interactive prompting patterns to design an optimal AI-powered question bank creation process. The generated questions are evaluated through a “Blind test” survey sent to various stakeholders including lecturers and learners. Initial results at the Banking Academy of Vietnam are relatively promising, suggesting a potential direction to streamline the time and effort involved in assessing learners at higher education institutes.
arXiv:2507.21174v2 Announce Type: replace-cross
Abstract: Large language models have been widely applied in many aspects of real life, bringing significant efficiency to businesses and offering distinctive user experiences. In this paper, we focus on exploring the application of ChatGPT, a chatbot based on a large language model, to support higher educator in generating quiz questions and assessing learners. Specifically, we explore interactive prompting patterns to design an optimal AI-powered question bank creation process. The generated questions are evaluated through a “Blind test” survey sent to various stakeholders including lecturers and learners. Initial results at the Banking Academy of Vietnam are relatively promising, suggesting a potential direction to streamline the time and effort involved in assessing learners at higher education institutes. Read More
KIX: A Knowledge and Interaction-Centric Metacognitive Framework for Task Generalizationcs.AI updates on arXiv.orgon July 31, 2025 at 4:00 am arXiv:2402.05346v3 Announce Type: replace
Abstract: People aptly exhibit general intelligence behaviors through flexible problem-solving and the ability to adapt to novel situations by reusing and applying high-level knowledge acquired over time. In contrast, artificial agents tend to be specialists, lacking such generalist behaviors. To bridge this gap, artificial agents will require understanding and exploiting critical structured knowledge representations. We introduce a metacognitive reasoning framework, Knowledge-Interaction-eXecution (KIX), and argue that interactions with objects, by leveraging a type space, facilitate the learning of transferable interaction concepts and promote generalization. This framework offers a principled approach for integrating knowledge into reinforcement learning and holds promise as an enabler for generalist behaviors in artificial intelligence, robotics, and autonomous systems.
arXiv:2402.05346v3 Announce Type: replace
Abstract: People aptly exhibit general intelligence behaviors through flexible problem-solving and the ability to adapt to novel situations by reusing and applying high-level knowledge acquired over time. In contrast, artificial agents tend to be specialists, lacking such generalist behaviors. To bridge this gap, artificial agents will require understanding and exploiting critical structured knowledge representations. We introduce a metacognitive reasoning framework, Knowledge-Interaction-eXecution (KIX), and argue that interactions with objects, by leveraging a type space, facilitate the learning of transferable interaction concepts and promote generalization. This framework offers a principled approach for integrating knowledge into reinforcement learning and holds promise as an enabler for generalist behaviors in artificial intelligence, robotics, and autonomous systems. Read More
The two people shaping the future of OpenAI’s researchMIT Technology Reviewon July 31, 2025 at 9:06 am For the past couple of years, OpenAI has felt like a one-man brand. With his showbiz style and fundraising glitz, CEO Sam Altman overshadows all other big names on the firm’s roster. Even his bungled ouster ended with him back on top—and more famous than ever. But look past the charismatic frontman and you get…
For the past couple of years, OpenAI has felt like a one-man brand. With his showbiz style and fundraising glitz, CEO Sam Altman overshadows all other big names on the firm’s roster. Even his bungled ouster ended with him back on top—and more famous than ever. But look past the charismatic frontman and you get… Read More
LLMs and Mental HealthTowards Data Scienceon July 31, 2025 at 3:01 pm Are LLMs good or bad for our mental health? It’s more complicated than that.
The post LLMs and Mental Health appeared first on Towards Data Science.
Are LLMs good or bad for our mental health? It’s more complicated than that.
The post LLMs and Mental Health appeared first on Towards Data Science. Read More
How to Benchmark LLMs – ARC AGI 3Towards Data Scienceon July 31, 2025 at 3:08 pm Learn how to LLMs are benchmarked, and try out the newly released ARC AGI 3
The post How to Benchmark LLMs – ARC AGI 3 appeared first on Towards Data Science.
Learn how to LLMs are benchmarked, and try out the newly released ARC AGI 3
The post How to Benchmark LLMs – ARC AGI 3 appeared first on Towards Data Science. Read More
The ONLY Data Science Roadmap You Need to Get a JobTowards Data Scienceon July 31, 2025 at 2:52 pm Are you looking to become a data scientist and don’t know where to start? In this article, I want to provide you with a straightforward, no-nonsense learning roadmap that you can follow to break into the industry. By the end, you’ll finally have a clear understanding of what is required and the best resources to
The post The ONLY Data Science Roadmap You Need to Get a Job appeared first on Towards Data Science.
Are you looking to become a data scientist and don’t know where to start? In this article, I want to provide you with a straightforward, no-nonsense learning roadmap that you can follow to break into the industry. By the end, you’ll finally have a clear understanding of what is required and the best resources to
The post The ONLY Data Science Roadmap You Need to Get a Job appeared first on Towards Data Science. Read More
Alibaba’s AI coding tool raises security concerns in the WestAI Newson July 30, 2025 at 10:00 am Alibaba has released a new AI coding model called Qwen3-Coder, built to handle complex software tasks using a large open-source model. The tool is part of Alibaba’s Qwen3 family and is being promoted as the company’s most advanced coding agent to date. The model uses a Mixture of Experts (MoE) approach, activating 35 billion parameters
The post Alibaba’s AI coding tool raises security concerns in the West appeared first on AI News.
Alibaba has released a new AI coding model called Qwen3-Coder, built to handle complex software tasks using a large open-source model. The tool is part of Alibaba’s Qwen3 family and is being promoted as the company’s most advanced coding agent to date. The model uses a Mixture of Experts (MoE) approach, activating 35 billion parameters
The post Alibaba’s AI coding tool raises security concerns in the West appeared first on AI News. Read More