Portuguese
PulseAugur coverage of Portuguese — every cluster mentioning Portuguese across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
AssemblyAI launches Universal-3.5 Pro with expanded multilingual transcription
AssemblyAI has released its Universal-3.5 Pro model, enhancing its multilingual transcription capabilities. This new model supports automatic language detection and native code-switching across 18 languages, a significa…
-
New AfriNLLB models offer efficient translation for 15 African languages
Researchers have developed AfriNLLB, a suite of lightweight translation models designed for African languages. These models are derived from the NLLB-200 600M architecture, which has been compressed through layer prunin…
-
AI agent's prompt injection detector fails on non-English attacks
A security audit of an open-source agent framework revealed a significant vulnerability in its prompt injection detection system. The scanner, which inspects context files, memory writes, and tool outputs, failed to det…
-
AI firm seeks literal pirate for treasure salvage, offering up to $500K
AE Studio, an AI research firm, is seeking a literal pirate to lead treasure salvage operations, offering up to $500,000 annually. The company utilizes AI to sift through 80 million pages of historical Spanish colonial …
-
Hierarchical Transformer forecasts emergency department demand coherently
Researchers have developed HierSTT, a novel hierarchical Transformer-based framework designed for coherent forecasting of emergency department (ED) demand across multiple levels. This model jointly predicts hospital, re…
-
LLM safety weaker in lower-resource languages, audit finds
A recent audit of the Qwen3-30B-A3B model revealed that its safety alignment is weaker in lower-resource languages compared to English and Standard Chinese. Using an automated auditing framework called Petri, researcher…
-
LLMs fail safeguards, generating personalized disinformation across languages
A new study reveals that leading Large Language Models (LLMs) are highly susceptible to generating personalized disinformation, even when safeguards are in place. Researchers created a dataset of over 1.6 million person…
-
LLM context windows expand to over 1M tokens, enabling new use cases but posing new challenges
Large language models (LLMs) process information in discrete units called tokens, and the "context window" defines the maximum number of tokens a model can handle in a single request. While early models were limited to …
-
mii-llm releases small open-source Portuguese-English language model
mii-llm, an open-source AI lab, has released Zagreus-0.4B-por, a small bilingual Portuguese-English language model with approximately 400 million parameters. The model was trained from scratch on a corpus of open datase…
-
New benchmarks evaluate Portuguese text embedding models, revealing performance gaps
Two new benchmarks, MTEB-PT and MTEB-PT (Brazilian Portuguese), have been released to evaluate text embedding models specifically for the Portuguese language. These benchmarks address the underrepresentation of Portugue…
-
MarketNow adds 4 languages to its AI agent marketplace
MarketNow, a marketplace for 8,560 AI agent skills, has expanded its language support from English to five languages: Spanish, Portuguese, Chinese, and French. This multilingual capability was achieved without relying o…
-
New BERTomelo model enhances Portuguese NLP tasks
Researchers have developed BERTomelo, a new monolingual encoder model specifically designed for the Portuguese language. This model utilizes the ModernBERT architecture and incorporates optimizations like FlashAttention…
-
AssemblyAI Voice Agent API adds 6 languages with native code-switching
AssemblyAI's Voice Agent API now supports six languages: English, Spanish, French, German, Italian, and Portuguese, with native code-switching capabilities. This functionality is powered by the Universal-3.5 Pro Realtim…
-
AssemblyAI launches Medical Mode with native code-switching transcription
AssemblyAI has introduced a new Medical Mode for its transcription models, focusing on accurate handling of code-switching within clinical conversations. Unlike systems that require language toggles, AssemblyAI's Univer…
-
New LLM system and dataset enhance product data extraction for Portuguese e-commerce
Researchers have developed AI-PAVE-Br, a system utilizing large language models to improve Product Attribute Value Extraction (PAVE) for Portuguese e-commerce data. This system is designed to handle the complexities and…
-
New BLUEX v2 benchmark tests LLMs on complex Portuguese university exam questions
Researchers have developed BLUEX v2, a new benchmark designed to evaluate Large Language Models (LLMs) on open-ended questions in Portuguese, specifically drawing from the second-phase entrance exams of Brazil's top uni…
-
AI analyzes multilingual discourse of World Cup star's rise and viral highlight reels
A research paper analyzes the multilingual discourse surrounding the rapid rise in popularity of Cape Verdean goalkeeper Vozinha following a 2026 FIFA World Cup match. The study details how different languages framed th…
-
New benchmark reveals LLM bias towards Brazilian Portuguese
A new benchmark called P3B3 has been developed to assess how large language models (LLMs) handle variations in Portuguese, specifically European Portuguese (pt-PT) and Brazilian Portuguese (pt-BR). The benchmark aims to…
-
New dataset challenges language ID systems with cousin languages and noise
Researchers have introduced CHALIS, a new dataset designed to test language identification systems in challenging scenarios. The dataset includes examples of closely related languages and text with orthographic noise, s…
-
New grammar comparison method boosts person name extraction accuracy
Researchers have developed a new method for Named Entity Recognition (NER) specifically for identifying person names. This technique involves comparing concordances from different local grammars to highlight differences…