The initiative marks the first city-led grant program supporting startups using AI to tackle civic issues

Overcome deployment barriers with a 'land and expand' approach that proves value early while maintaining compliance in financial services

QubitCast system aims to detect extreme weather events up to six months in advance while using fewer computational resources

Adoption is growing fast, but knowledge gaps, trust issues and the need for reskilling threaten to hold industries back.

Also inside, AI helps fans supercharge their sports experience; safeguards launched for AI shopping agents, and more

The Multiverse Computing nano model packs conversational intelligence space 15,000 times smaller than a chicken brain

New survey finds only a fraction of companies complete complex data migrations on time and smarter AI use may be the missing piece

Traditional information systems are becoming the essential foundation for next-generation AI solutions

Seed-Prover: Deep and Broad Reasoning for Automated Theorem Proving

Collaboration aims to integrate quantum computers with AI accelerators and high-performance computing systems to solve previously intractable problems

New framework introduces tamper-proof digital credentials to validate AI agents and reduce instances of fraud

Why tomorrow’s top planners will prepare faster, decide smarter and drive broader impact

The Vayu acquisition will bolster Serve’s physical AI capabilities

Related Articles

• Apple Begins Breaking Up Its AI Team With Robotics, Siri Changes - Bloomberg.com
• Navigating the AI and Robotics revolution: Courses for students to pursue - India Today
• Geekplus named Top 50 robotics company in the world for the fourth time

Apple Begins Breaking Up Its AI Team With Robotics, Siri Changes  Bloomberg.com

Navigating the AI and Robotics revolution: Courses for students to pursue  India Today

Geekplus, the global leader in warehouse robotics, proudly announces its fourth recognition in the prestigious 2025 RBR50 Innovation Awards by the Robotics Business Review (RBR), standing alongside global tech giants like ABB and Nvidia the forefront of innovation. Regarded as the “Oscars” of the robotics industry, the RBR50 Innovation Awards...The post Geekplus named Top 50 robotics company in the world for the fourth time first appeared on AI-Tech Park.

The letter addresses CEOs of OpenAI, Meta, Microsoft, Apple, Google, and Elon Musk’s xAI.

Related Articles

• When OpenAI Hired an Illusionist and a Magician to Build ChatGPT
• OpenAI’s New Model is More Human Than Humans, Study Finds
• OpenAI is editing its GPT-5 rollout on the fly — here’s what’s changing in ChatGPT

Building a shark camouflage suit led Andrew Mayne down the path of AI.The post When OpenAI Hired an Illusionist and a Magician to Build ChatGPT appeared first on Analytics India Magazine.

OpenAI’s new model has passed a version of the Turing test, with participants judging the AI chatbot as human 73% of the time

OpenAI must stabilize infrastructure, tune personalization, and decide how to moderate immersive interactions.

WeRide’s AiDrive will be integrated into passenger vehicles with mass production, deployment slated for later this year

xAI is launching a federal lawsuit against Apple and OpenAI, claiming their exclusive ChatGPT integration on iPhones creates a monopoly that sidelines rival AI chatbots.

Related Articles

• MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Speculative Decoding
• When OpenAI Hired an Illusionist and a Magician to Build ChatGPT
• OpenAI’s New Model is More Human Than Humans, Study Finds

arXiv:2408.11049v4 Announce Type: replace Abstract: Large Language Models (LLMs) have become more prevalent in long-context applications such as interactive chatbots, document analysis, and agent workflows, but it is challenging to serve long-context requests with low latency and high throughput. Speculative decoding (SD) is a widely used technique to reduce latency losslessly, but the conventional wisdom suggests that its efficacy is limited to small batch sizes. In MagicDec, we show that surprisingly SD can achieve speedup even for a high throughput inference regime for moderate to long sequences. More interestingly, an intelligent drafting strategy can achieve better speedup with increasing batch size based on our rigorous analysis. MagicDec first identifies the bottleneck shifts with increasing batch size and sequence length, and uses these insights to deploy SD more effectively for high throughput inference. We leverage draft model with sparse KV cache to address the KV bottleneck, which scales with both sequence length and batch size. Additionally, we propose a theoretical model to select the optimal drafting strategy for maximum speedup. Our work highlights the broad applicability of speculative decoding in long-context serving, as it can enhance throughput and reduce latency without compromising accuracy. For moderate to long sequences, we demonstrate up to 2.51x speedup for Llama3.1-8B when serving batch sizes ranging from 32 to 256 on various types of hardware and tasks.

Building a shark camouflage suit led Andrew Mayne down the path of AI.The post When OpenAI Hired an Illusionist and a Magician to Build ChatGPT appeared first on Analytics India Magazine.

OpenAI’s new model has passed a version of the Turing test, with participants judging the AI chatbot as human 73% of the time

From real-time match updates to personalized highlights, fans are demanding smarter, faster and more connected ways to follow the action

Seemplicity will use the funds to accelerate AI innovation and expand globally

OpenAI is updating its latest GPT model following user complaints

Related Articles

• When OpenAI Hired an Illusionist and a Magician to Build ChatGPT
• OpenAI’s New Model is More Human Than Humans, Study Finds
• OpenAI is editing its GPT-5 rollout on the fly — here’s what’s changing in ChatGPT

Building a shark camouflage suit led Andrew Mayne down the path of AI.The post When OpenAI Hired an Illusionist and a Magician to Build ChatGPT appeared first on Analytics India Magazine.

OpenAI’s new model has passed a version of the Turing test, with participants judging the AI chatbot as human 73% of the time

OpenAI must stabilize infrastructure, tune personalization, and decide how to moderate immersive interactions.

Market volatility and unexpected tariffs are overwhelming traditional systems, but AI offers pattern recognition and adaptability when it matters most

Related Articles

• From Patterns to Context: How AI in Medical Imaging Is Evolving from Pattern Recognition to Clinical Reasoning - DataDrivenInvestor
• Preventing DeFi Hack Events with Chainalysis Hexagate Using Pattern Recognition and Machine Learning - Chainalysis
• ParaScript Launches New AI-Powered Check Stock Verification Solution

From Patterns to Context: How AI in Medical Imaging Is Evolving from Pattern Recognition to Clinical Reasoning  DataDrivenInvestor

Preventing DeFi Hack Events with Chainalysis Hexagate Using Pattern Recognition and Machine Learning  Chainalysis

ParaScript, an AI-powered document processing company, today announced the launch of a new check stock verification solution, CheckStock.AI®, an enhancement of the company’s current check fraud detection capabilities. The AI-powered technology uses a combination of image analysis and pattern recognition to accurately detect counterfeit checks, alterations, or the use of unauthorized...The post ParaScript Launches New AI-Powered Check Stock Verification Solution first appeared on AI-Tech Park.

Also inside, Oracle integrates OpenAI's GPT-5 across enterprise software portfolio; Intel shares surge as SoftBank invests $2B amid Trump admin stake plans, and more

Related Articles

• When OpenAI Hired an Illusionist and a Magician to Build ChatGPT
• EBU News Report 2025: Leading Newsrooms in the Age of Generative AI - EBU
• GOLFMANAGER LAUNCHES THE FIRST PRACTICAL PROMPTING GUIDE TO HELP GOLF CLUB MANAGERS HARNESS THE POWER OF GENERATIVE AI - The Golf Wire

Building a shark camouflage suit led Andrew Mayne down the path of AI.The post When OpenAI Hired an Illusionist and a Magician to Build ChatGPT appeared first on Analytics India Magazine.

EBU News Report 2025: Leading Newsrooms in the Age of Generative AI  EBU

GOLFMANAGER LAUNCHES THE FIRST PRACTICAL PROMPTING GUIDE TO HELP GOLF CLUB MANAGERS HARNESS THE POWER OF GENERATIVE AI  The Golf Wire

What if, one day, AI models do develop something like subjective experience?

Deal with Helm.ai focuses on Advanced Driver Assistance Systems for Honda’s mainstream vehicles

Related Articles

• Evaluation of (Un-)Supervised Machine Learning Methods for GNSS Interference Classification with Real-World Data Discrepancies
• UN Warns of Terrorist Threat for Self-Driving Cars, Slaughterbots
• Nissan Self-Driving Cars to Use AI Tech From British Startup

arXiv:2503.23775v1 Announce Type: new Abstract: The accuracy and reliability of vehicle localization on roads are crucial for applications such as self-driving cars, toll systems, and digital tachographs. To achieve accurate positioning, vehicles typically use global navigation satellite system (GNSS) receivers to validate their absolute positions. However, GNSS-based positioning can be compromised by interference signals, necessitating the identification, classification, determination of purpose, and localization of such interference to mitigate or eliminate it. Recent approaches based on machine learning (ML) have shown superior performance in monitoring interference. However, their feasibility in real-world applications and environments has yet to be assessed. Effective implementation of ML techniques requires training datasets that incorporate realistic interference signals, including real-world noise and potential multipath effects that may occur between transmitter, receiver, and satellite in the operational area. Additionally, these datasets require reference labels. Creating such datasets is often challenging due to legal restrictions, as causing interference to GNSS sources is strictly prohibited. Consequently, the performance of ML-based methods in practical applications remains unclear. To address this gap, we describe a series of large-scale measurement campaigns conducted in real-world settings at two highway locations in Germany and the Seetal Alps in Austria, and in large-scale controlled indoor environments. We evaluate the latest supervised ML-based methods to report on their performance in real-world settings and present the applicability of pseudo-labeling for unsupervised learning. We demonstrate the challenges of combining datasets due to data discrepancies and evaluate outlier detection, domain adaptation, and data augmentation techniques to present the models' capabilities to adapt to changes in the datasets.

A new report examines what the U.N. crime institute director describes as 'the dark side of AI’

The Japanese automaker will use Wayve’s embodied AI tech for its next-generation driver assistance features

Musk's bold $97.6 billion bid to acquire OpenAI signals a seismic shift in the AI landscape. Despite the deal's collapse, Musk swiftly pivoted to launch xAI, while Meta.

Related Articles

• When OpenAI Hired an Illusionist and a Magician to Build ChatGPT
• OpenAI’s New Model is More Human Than Humans, Study Finds
• OpenAI is editing its GPT-5 rollout on the fly — here’s what’s changing in ChatGPT

Building a shark camouflage suit led Andrew Mayne down the path of AI.The post When OpenAI Hired an Illusionist and a Magician to Build ChatGPT appeared first on Analytics India Magazine.

OpenAI’s new model has passed a version of the Turing test, with participants judging the AI chatbot as human 73% of the time

OpenAI must stabilize infrastructure, tune personalization, and decide how to moderate immersive interactions.

X-Omni: Reinforcement Learning Makes Discrete Autoregressive Image Generative Models Great Again

Related Articles

• Optimizing Case-Based Reasoning System for Functional Test Script Generation with Large Language Models
• Understanding R1-Zero-Like Training: A Critical Perspective
• Nuclear Microreactor Control with Deep Reinforcement Learning

arXiv:2503.20576v1 Announce Type: cross Abstract: In this work, we explore the potential of large language models (LLMs) for generating functional test scripts, which necessitates understanding the dynamically evolving code structure of the target software. To achieve this, we propose a case-based reasoning (CBR) system utilizing a 4R cycle (i.e., retrieve, reuse, revise, and retain), which maintains and leverages a case bank of test intent descriptions and corresponding test scripts to facilitate LLMs for test script generation. To improve user experience further, we introduce Re4, an optimization method for the CBR system, comprising reranking-based retrieval finetuning and reinforced reuse finetuning. Specifically, we first identify positive examples with high semantic and script similarity, providing reliable pseudo-labels for finetuning the retriever model without costly labeling. Then, we apply supervised finetuning, followed by a reinforcement learning finetuning stage, to align LLMs with our production scenarios, ensuring the faithful reuse of retrieved cases. Extensive experimental results on two product development units from Huawei Datacom demonstrate the superiority of the proposed CBR+Re4. Notably, we also show that the proposed Re4 method can help alleviate the repetitive generation issues with LLMs.

arXiv:2503.20783v1 Announce Type: cross Abstract: DeepSeek-R1-Zero has shown that reinforcement learning (RL) at scale can directly enhance the reasoning capabilities of LLMs without supervised fine-tuning. In this work, we critically examine R1-Zero-like training by analyzing its two core components: base models and RL. We investigate a wide range of base models, including DeepSeek-V3-Base, to understand how pretraining characteristics influence RL performance. Our analysis reveals that DeepSeek-V3-Base already exhibit ''Aha moment'', while Qwen2.5 base models demonstrate strong reasoning capabilities even without prompt templates, suggesting potential pretraining biases. Additionally, we identify an optimization bias in Group Relative Policy Optimization (GRPO), which artificially increases response length (especially for incorrect outputs) during training. To address this, we introduce Dr. GRPO, an unbiased optimization method that improves token efficiency while maintaining reasoning performance. Leveraging these insights, we present a minimalist R1-Zero recipe that achieves 43.3% accuracy on AIME 2024 with a 7B base model, establishing a new state-of-the-art. Our code is available at https://github.com/sail-sg/understand-r1-zero.

arXiv:2504.00156v1 Announce Type: cross Abstract: The economic feasibility of nuclear microreactors will depend on minimizing operating costs through advancements in autonomous control, especially when these microreactors are operating alongside other types of energy systems (e.g., renewable energy). This study explores the application of deep reinforcement learning (RL) for real-time drum control in microreactors, exploring performance in regard to load-following scenarios. By leveraging a point kinetics model with thermal and xenon feedback, we first establish a baseline using a single-output RL agent, then compare it against a traditional proportional-integral-derivative (PID) controller. This study demonstrates that RL controllers, including both single- and multi-agent RL (MARL) frameworks, can achieve similar or even superior load-following performance as traditional PID control across a range of load-following scenarios. In short transients, the RL agent was able to reduce the tracking error rate in comparison to PID. Over extended 300-minute load-following scenarios in which xenon feedback becomes a dominant factor, PID maintained better accuracy, but RL still remained within a 1% error margin despite being trained only on short-duration scenarios. This highlights RL's strong ability to generalize and extrapolate to longer, more complex transients, affording substantial reductions in training costs and reduced overfitting. Furthermore, when control was extended to multiple drums, MARL enabled independent drum control as well as maintained reactor symmetry constraints without sacrificing performance -- an objective that standard single-agent RL could not learn. We also found that, as increasing levels of Gaussian noise were added to the power measurements, the RL controllers were able to maintain lower error rates than PID, and to do so with less control effort.

Fans will have access to tools including an interactive chatbot for match insights, as well as AI commentary and summarization

The new feature comes as part of Anthropic’s push for better AI welfare

The new competition is being launched under Gates’ Alzheimer’s Disease Data Initiative

Beyond the hype, mature AI systems now deliver real-time decision-making capabilities that revolutionize operations in complex sectors

Open-source Surya model analyzes solar data to forecast space weather impacts on Earth's critical infrastructure

Head to the Tools dropdown menu and select Audio > Listen to this tab to enable it.

Japanese tech giant and US government show confidence in chipmaker's future despite recent workforce cuts and public criticism from Trump

Latest AI model brings advanced reasoning and automation capabilities to Oracle's database and cloud applications

Related Articles

• When OpenAI Hired an Illusionist and a Magician to Build ChatGPT
• OpenAI’s New Model is More Human Than Humans, Study Finds
• OpenAI is editing its GPT-5 rollout on the fly — here’s what’s changing in ChatGPT

Building a shark camouflage suit led Andrew Mayne down the path of AI.The post When OpenAI Hired an Illusionist and a Magician to Build ChatGPT appeared first on Analytics India Magazine.

OpenAI’s new model has passed a version of the Turing test, with participants judging the AI chatbot as human 73% of the time

OpenAI must stabilize infrastructure, tune personalization, and decide how to moderate immersive interactions.

The new tool could help astronauts diagnose and treat medical issues without real-time contact with Earth

Insilico Medicine said its novel medication could be a “paradigm shift” in Parkinson’s treatment

Related Articles

• EBU News Report 2025: Leading Newsrooms in the Age of Generative AI - EBU
• GOLFMANAGER LAUNCHES THE FIRST PRACTICAL PROMPTING GUIDE TO HELP GOLF CLUB MANAGERS HARNESS THE POWER OF GENERATIVE AI - The Golf Wire
• Movies Made Using Generative AI Can Now Win Oscars - PCMag UK

EBU News Report 2025: Leading Newsrooms in the Age of Generative AI  EBU

GOLFMANAGER LAUNCHES THE FIRST PRACTICAL PROMPTING GUIDE TO HELP GOLF CLUB MANAGERS HARNESS THE POWER OF GENERATIVE AI  The Golf Wire

Movies Made Using Generative AI Can Now Win Oscars  PCMag UK

Strong governance frameworks ensure AI agents deliver value while maintaining human trust

It will only activate in “rare, extreme cases” when users repeatedly push the AI toward harmful or abusive topics.

The document showed an AI responding to a high school student’s romantic prompt with an affectionate, intimate message.

China’s leading artificial intelligence startup DeepSeek Ltd., which took the industry by storm earlier this year, was reportedly forced to delay the release of its upcoming R2 model after struggling to train it using chips supplied by Huawei Technologies Co. Ltd. The delay highlights the ongoing struggles China faces as it pushes domestic AI companies […]The post DeepSeek R2 model release reportedly held back by faulty Huawei chips appeared first on SiliconANGLE.

Related Articles

• Deloitte clears a faster AI path for banks with a governed data platform
• AI pioneer Yoshua Bengio launches nonprofit LawZero AI lab, with a focus on safe AI
• Space industry confronts twin disruptors: AI and geopolitics

Financial institutions are rapidly moving to reimagine their operations with generative artificial intelligence, which begins with a governed data platform. But these organizations aren’t chasing hype: They’re demanding measurable value. Deloitte Consulting LLP’s data leaders say clients want more than strategies — they expect accelerated delivery, real-time insight and a measurable return on investment. Deloitte’s […]The post Deloitte clears a faster AI path for banks with a governed data platform appeared first on SiliconANGLE.

Yoshua Bengio, one of the world’s most prominent computer scientists, is launching a new artificial intelligence lab called LawZero. Bengio told the Financial Times on Monday that the group will seek to develop safety-optimized AI systems. The effort is backed by about $30 million in donations from a who’s who of backers. The list includes […]The post AI pioneer Yoshua Bengio launches nonprofit LawZero AI lab, with a focus on safe AI appeared first on SiliconANGLE.

Space businesses are under pressure to adapt as artificial intelligence and shifting geopolitics reshape their industry in ways that are still coming into focus.The post Space industry confronts twin disruptors: AI and geopolitics appeared first on SpaceNews.

Google LLC’s DeepMind artificial intelligence lab has released one of its smallest models yet in the shape of Gemma 3 270M, with just 270 million parameters. That means it’s much smaller than many of the most powerful frontier large language models, which generally have billions of parameters, or internal settings that govern their behavior. The […]The post Google’s Gemma 3 270M is a compact yet powerful AI model that can run on your toaster appeared first on SiliconANGLE.

Related Articles

• MagicDec: Breaking the Latency-Throughput Tradeoff for Long Context Generation with Speculative Decoding
• High-Dimension Human Value Representation in Large Language Models
• QualiSpeech: A Speech Quality Assessment Dataset with Natural Language Reasoning and Descriptions

arXiv:2408.11049v4 Announce Type: replace Abstract: Large Language Models (LLMs) have become more prevalent in long-context applications such as interactive chatbots, document analysis, and agent workflows, but it is challenging to serve long-context requests with low latency and high throughput. Speculative decoding (SD) is a widely used technique to reduce latency losslessly, but the conventional wisdom suggests that its efficacy is limited to small batch sizes. In MagicDec, we show that surprisingly SD can achieve speedup even for a high throughput inference regime for moderate to long sequences. More interestingly, an intelligent drafting strategy can achieve better speedup with increasing batch size based on our rigorous analysis. MagicDec first identifies the bottleneck shifts with increasing batch size and sequence length, and uses these insights to deploy SD more effectively for high throughput inference. We leverage draft model with sparse KV cache to address the KV bottleneck, which scales with both sequence length and batch size. Additionally, we propose a theoretical model to select the optimal drafting strategy for maximum speedup. Our work highlights the broad applicability of speculative decoding in long-context serving, as it can enhance throughput and reduce latency without compromising accuracy. For moderate to long sequences, we demonstrate up to 2.51x speedup for Llama3.1-8B when serving batch sizes ranging from 32 to 256 on various types of hardware and tasks.

arXiv:2404.07900v4 Announce Type: replace Abstract: The widespread application of LLMs across various tasks and fields has necessitated the alignment of these models with human values and preferences. Given various approaches of human value alignment, there is an urgent need to understand the scope and nature of human values injected into these LLMs before their deployment and adoption. We propose UniVaR, a high-dimensional neural representation of symbolic human value distributions in LLMs, orthogonal to model architecture and training data. This is a continuous and scalable representation, self-supervised from the value-relevant output of 8 LLMs and evaluated on 15 open-source and commercial LLMs. Through UniVaR, we visualize and explore how LLMs prioritize different values in 25 languages and cultures, shedding light on complex interplay between human values and language modeling.

arXiv:2503.20290v1 Announce Type: cross Abstract: This paper explores a novel perspective to speech quality assessment by leveraging natural language descriptions, offering richer, more nuanced insights than traditional numerical scoring methods. Natural language feedback provides instructive recommendations and detailed evaluations, yet existing datasets lack the comprehensive annotations needed for this approach. To bridge this gap, we introduce QualiSpeech, a comprehensive low-level speech quality assessment dataset encompassing 11 key aspects and detailed natural language comments that include reasoning and contextual insights. Additionally, we propose the QualiSpeech Benchmark to evaluate the low-level speech understanding capabilities of auditory large language models (LLMs). Experimental results demonstrate that finetuned auditory LLMs can reliably generate detailed descriptions of noise and distortion, effectively identifying their types and temporal characteristics. The results further highlight the potential for incorporating reasoning to enhance the accuracy and reliability of quality assessments. The dataset will be released at https://huggingface.co/datasets/tsinghua-ee/QualiSpeech.