Results for “Prompt”

The Guardian US Ramon Antonio Vargas

Bomb found at Illinois ethanol plant, prompting evacuations

Peoria officials said a bomb was found at BioUrja Renewables and later defused; FBI said it is investigating Residents were ordered to evacuate parts of an Illinois community on Friday after the mayor there said authorities found a bomb at a local ethanol- and alcohol-production plant. The discovery reported at BioUrja Renewables in Peoria triggered a mandatory evacuation over a one-mile evacuation zone encompassing sections of that city as well as the adjacent municipality…

Read at The Guardian US

Tehran Times (Iranian state-aligned)

Isfahan’s pigeon towers emerge as potential World Heritage candidate

TEHRAN – Iran’s historic pigeon towers in Isfahan province could have the potential to be nominated for UNESCO World Heritage status, prompting renewed calls for comprehensive documentation, comparative research and stronger conservation measures to establish their significance on a global scale.

Read at Tehran Times (Iranian state-aligned)

Khaama Press Madina Hasan

ICC Suspends Part of Afghanistan War Crimes Investigation

The International Criminal Court (ICC) has suspended part of its investigation into alleged war crimes in Afghanistan, including accusations against US forces and Islamic State Khorasan (IS-K), prompting an appeal from the court’s prosecution. The judges’ decision concerns a case involving allegations against US forces and IS-K. The partial suspension of the investigation was made

Read at Khaama Press

DW Geopolitics

Nigeria: How the pcoming elections are influeced by the rise of AI disinformation

As Nigeria inches closer to the 2027 polls, the use of artificial intelligence in the electoral process has prompted questions about integrity and institutional readiness. Can Nigerians trust their democratic processes?

Read at DW

arXiv (research papers) Pengyu Chen SI

Mem2Fair: Online Fairness Control for Text-to-Video Generation via Fair Memory

Text-to-video (T2V) generation models can produce systematically unfair demographic outcomes when prompts leave protected attributes unspecified. For example, prompts describing nurses without specifying gender may disproportionately generate videos depicting female nurses. Mitigating unfairness in T2V generation is particularly challenging: training-based methods that modify model internals require architecture-specific access and risk disrupting temporal consistency, while…

Read at arXiv (research papers)

arXiv (research papers) Nikhil Kumar Jha SI

What Moves a LoCoMo Score? Memory Context, Answer Prompts and Judges

LoCoMo is a widely reported benchmark for long-term conversational memory, but published scores mix memory quality with choices of answer prompt, judge and baseline layout that are rarely reported together. We measure how much these choices move a score, and evaluate Becos, a team memory system for coding agents, as a black box under all of them: ten conversations (1,540 questions), two answer models (gpt-4o-mini, gpt-4.1-mini), two answer prompts and three LLM judges…

Read at arXiv (research papers)

El País English Beatriz Guillén ,Patricia Clarembaux Geopolitics

The migrant caravan crossing Mexico that has triggered Trump’s new tariff threat: ‘Our fear is that the Mexican president will send us back’

About 1,500 people are walking from the country’s south toward the United States, prompting Donald Trump to again threaten Claudia Sheinbaum with doubling trade tariffs

Read at El País English

arXiv (research papers) Charan Teja Nagisettygari SI

How Much Determinism Can Fine-Tuning Buy? Margin Regularisation Against Batch-Induced Token Flips

Greedy decoding at zero temperature is widely assumed to be deterministic, although the exact same prompt may yield different tokens depending on the number of other prompts sharing the batch for inference. The currently available solutions operate during inference, trading determinism for throughput. Can we train a model in such a way that fewer decoding steps would be susceptible? We characterize the perturbation empirically across six model families and three numerical…

Read at arXiv (research papers)

arXiv (research papers) Xiaotang Gai SI

DecoEdit: Beyond Flat Control for 3D-Aware Unified Image Editing

While recent advances in multimodal image editing have achieved impressive results, current frameworks remain largely confined to ``flat" 2D controls (e.g., text prompts, bounding boxes, or masks). Precise 3D-aware geometric manipulation -- such as coordinating an object's location, physical scale, and 3D orientation while strictly preserving its fine-grained identity -- remains an elusive challenge, typically suffering from severe detail degradation or requiring expensive…

Read at arXiv (research papers)

South China Morning Post Resty Woro Yuniar Geopolitics

Indonesia’s commodity export U-turn: no monopoly, just all the data

Indonesia has abandoned plans to make a powerful new state-owned enterprise the “sole exporter” of its palm oil, coal and ferroalloys, in a move that has reassured some businesses, even as questions over oversight and pricing remain. President Prabowo Subianto’s original proposal to tighten the government’s control over the three major commodities, put forward in May, had prompted speculation that state-linked companies could gain an unfair advantage through access to…

Read at South China Morning Post

arXiv (research papers) Adriana Watson Security

When the Spec Sheet Lies: Injection Attacks on LLM-Generated Compliance Artifacts

As Large Language Models (LLMs) become used more commonly across a variety of fields, a growing body of research has highlighted particular concerns surrounding model security. Notably, prompt injection attacks, where malicious instructions are inserted into prompts, are particularly simple to implement yet capable of inducing devastating effects on the fidelity of model outputs. While there is substantial research surrounding attack models and their effects on chat-based…

Read at arXiv (research papers)

NBC News

All accounted for after tanker crashes into New Orleans riverboat

More than 100 passengers and crew were rescued after a tanker struck a river cruise on the Mississippi River in New Orleans, prompting a massive rescue operation that sent 32 people to the hospital. NBC News’ George Solis reports.

Read at NBC News

Ukrinform

Evacuees from Oleshky promptly issued documents and bank cards – Lubinets

Residents evacuated from Oleshky were promptly issued documents and bank cards while staying at the Kyiv Regional Hospital. They are already receiving assistance from the Ukrainian government and international partners.

Read at Ukrinform

NBC News Aria Bendix

U.S. measles deaths rise to 10 after new reports in Pennsylvania and Wisconsin

Total measles deaths in the United States this year rose to 10 on Friday, a grim milestone in an outbreak that has seen a record number of infections and prompted state and local health departments to ramp up vaccination efforts.

Read at NBC News

Ukrainska Pravda Tetyana Oliynyk

Flights grounded across US due to communications system failure

Nearly 7,000 flights were delayed or cancelled across the United States on 13 November after a communications failure affected several air traffic control centres, prompting a nationwide ground stop.

Read at Ukrainska Pravda

NBC News Shayna Jacobs

Rescue underway after an oil tanker collides with a New Orleans river boat

A river boat with over 100 passengers was struck on Friday by a much larger tanker ship, capsizing the cruise vessel and prompting a major rescue effort in the Mississippi River near the French Quarter.

Read at NBC News

The Guardian World Ramon Antonio Vargas in New Orleans

Tanker collides with cruise ship on Mississippi River in New Orleans

A large emergency response is underway to rescue people from the water, with overnight cruise ship on its side A tanker collided with an overnight cruise ship on the Mississippi River in New Orleans on Friday afternoon, prompting a large emergency response, according to officials. The boat that was struck was American Cruise Lines’ American Heritage, a New Orleans port official told local news outlet WVUE. The ship had about 129 passengers – in addition to crew members…

Read at The Guardian World

The Guardian World Caroline Davies

Walliams allegations prompt calls for staggered age of consent at 16 and 17

Barrister say law does not adequately recognise power imbalance between young people and much older adults Allegations against Russell Brand and, separately, David Walliams, which he denies, have led to calls for greater legal protection for 16- and 17-year-olds from sexual exploitation by older adults. A staggered consent law is one route favoured by campaigners, supported by some Labour MPs. Claire Waxman, the victims’ commissioner for England and Wales, has called for…

Read at The Guardian World

arXiv (research papers) Wang Ma SI

Black-Box Uncertainty Quantification for Large Language Models via Ensemble-of-Ensembles

Reliable use of large language models (LLMs) requires uncertainty estimates that separate ambiguity intrinsic to a prompt from genuine knowledge gaps. Bayesian and deep-ensemble methods deliver this aleatoric/epistemic decomposition in principle, but are computationally prohibitive at LLM scale and require white-box access. Existing black-box methods, in contrast, summarize to a single scalar that cannot make the distinction. We introduce a two-level ensemble that bridges…

Read at arXiv (research papers)

arXiv (research papers) Anthony Abou Mrad Tech policy

When evaluation becomes intervention in clinical AI

AI judges, models that grade another system's output against a rubric, extend clinical review to the scale at which clinical AI writes. When teams use judge assessments to select a prompt or train a model, the judge becomes part of the intervention. Selection can then favor errors the judge overlooks. We describe what a judge's report must contain and the independent clinical test each judge-driven change needs.

Read at arXiv (research papers)

arXiv (research papers) Jonathan Pan Security

From Bytes to Behavior: Request-Level Attribution from Live GPU Hidden States

Existing memory forensics for machine learning systems typically targets static artifacts, such as model weights or prompt logs, rather than the transient computational states of a live inference process. This gap limits an examiner's ability to analyze the specific behavior of a running large language model (LLM) service that cannot be restarted or modified. We present a method for acquiring per-layer residual-stream activations from a live GPU process using a non-modifying…

Read at arXiv (research papers)

VietnamPlus (Vietnamese state media)

IMF sees positive growth prospects for Vietnam, urges cautious monetary policy

Martin Sommer, head of the IMF’s Article IV consultation mission team to Vietnam, noted that inflation risks in Vietnam remain under control, but several factors could add to price pressures. He therefore called for cautious monetary policy management and readiness to respond promptly if inflationary pressures increased. Vietnam strengthens financial cooperation at IMF-World Bank annual meetings Deputy PM receives IMF Asia-Pacific Director Vietnamese PM meets with leaders of…

Read at VietnamPlus (Vietnamese state media)

Al-Monitor

At least nine detained protesting at blast-hit Turkey steel plant

Turkey's security forces on Friday detained at least nine people protesting a ban on bringing mobile phones into a steel plant that was hit by a deadly explosion this week, a workers' collective said. Workers arriving for their morning shift at the HABAS steel factory were told to leave their mobile phones at a security checkpoint, prompting them to gather at the facility's entrance in protest, according to a spokesman for the Aegean Workers' Collective.

Read at Al-Monitor

South China Morning Post Chong Ming Lee,Minxiao Chang Geopolitics

Will Jensen Huang’s downfall tip the scales against open-source AI?

Jensen Huang’s abrupt resignation from Nvidia has cast uncertainty over the US chip giant’s multibillion-dollar investments in open-source artificial intelligence, while potentially depriving Chinese developers of one of their most influential advocates in Washington. His departure could prompt Nvidia to scale back spending on open-weight models and their supporting infrastructure, analysts said, although strong commercial incentives would likely keep the company committed…

Read at South China Morning Post

VietnamPlus (Vietnamese state media) Geopolitics

Vietnam always treasures traditional friendship with Cuba: PM

Prime Minister Le Minh Hung called on both sides to regularly review and promote the implementation of high-level agreements, and coordinate promptly in addressing emerging issues to ensure practical benefits for both countries. Vietnam joins international community in calling for end to US embargo against Cuba Vietnam, Cuba reaffirm enduring friendship, solidarity Vienna gathering helps strengthen Vietnam - Laos - Cuba friendship, solidarity

Read at VietnamPlus (Vietnamese state media)

Viet Nam News (state-run daily)

AI-powered People’s Assistant application streamlines administrative procedures

The system must be connected to the city police’s Information Security Monitoring Centre, while administrative forms, application requirements and procedures on the National Public Service Portal must be updated promptly.

Read at Viet Nam News (state-run daily)

Tech in Asia Naomi Li Gan Startups

OpenAI backs US push to restrict China’s access to AI chips

Existing US controls have prompted Nvidia to modify H20 chips for China.

Read at Tech in Asia

SiliconANGLE AI Ryan Stevens SI

Kubernetes gets a second act as AI agents turn one prompt into thousands of inference calls

The rise of agentic AI systems is giving Kubernetes, long valued for its reliability and predictability, an unexpected second act. As AI spending shifts from training models toward serving them and running agents, the open-source community is developing tools to handle the growing volume of inference requests. The orchestration platform was built from the start

Read at SiliconANGLE AI

AWS Machine Learning Blog Sai Pavan Kasani SI

Four Amazon Bedrock inference choices that affect cost

Your Amazon Bedrock inference bill can keep climbing even after you tune prompts and application logic. Four request-level choices (cache duration, service tier, reasoning effort, and maxTokens) change how many tokens the service processes and the rate applied. We show how we tested each and how to verify the effect.

Read at AWS Machine Learning Blog

Older stories