Czech
PulseAugur coverage of Czech — every cluster mentioning Czech across labs, papers, and developer communities, ranked by signal.
- instance of Serbian 90%
- instance of Croatian 90%
- instance of alphaXiv 70%
- used by alphaXiv 70%
- instance of Romanian 70%
- instance of Swedish 70%
- used by ScienceCast 70%
- instance of ScienceCast 70%
- instance of CatalyzeX 70%
- instance of Polish 70%
- instance of Slovak 70%
- instance of Consumer Electronics Show 60%
5 day(s) with sentiment data
-
AI translation errors found across multiple languages on dev.to sites
A software development team discovered over a thousand mistranslations across their websites, where words from one language appeared within sentences of another. These errors, generated by automated translation models, …
-
New multilingual datasets aim to boost AI in cardiology
Researchers have developed new multilingual datasets, CliTe and ParaCliTe, to advance AI in cardiology. These corpora consist of artificial clinical records generated by cardiologists across seven European languages: Cz…
-
CardioLM: Multilingual LLMs for Cardiology Domain
Researchers have developed CardioLM, a suite of multilingual small language models specifically designed for cardiology domain tasks. The project includes CardioCCC, a corpus of translated and annotated case reports in …
-
Machinarium Dev Criticizes AI-Generated Projects Using Game Assets
Amanita Design, the creators of the game Machinarium, have criticized the use of their game's assets in an AI-generated project. The studio expressed their disapproval, suggesting that creative works, like love letters,…
-
New study reveals "polarity blindness" vulnerability in LLMs
A new study published on arXiv investigates a vulnerability in large language models (LLMs) termed "polarity blindness." This weakness allows harmful intent to go unrecognized when linguistic transformations like negati…
-
Personalized ASR system achieves 50% error reduction for dysarthric speaker
Researchers have developed a personalized automatic speech recognition (ASR) system for a Czech speaker with dysarthria and a tracheostoma, whose speech is otherwise unintelligible. The system utilizes a multi-stage tra…
-
Compact HTML-LM Model Achieves State-of-the-Art for Czech Web Documents
Researchers have developed HTML-LM, a compact foundation model with 154 million parameters designed to efficiently process Czech HTML documents. This model, built on a ModernBERT architecture, utilizes HTML-aware traini…
-
Linguardia platform criticized for heavy reliance on AI and machine translation
A user on Mastodon has raised concerns about the language learning platform Linguardia, alleging significant use of generative AI and machine translation in its content. The user claims that the platform relies heavily …
-
Czech expert urges firms to prioritize process improvement over AI adoption
A Czech automation expert is urging local companies to prioritize process improvement over the immediate adoption of AI technologies. While Czech firms are actively investing in AI, this Prague-based specialist believes…
-
New method quantifies consonant contributions to word intelligibility
Researchers have developed a novel method to quantify the contribution of consonants to word intelligibility using acoustic masking. This technique involves masking individual consonants within a word and assessing whet…
-
LLMs show weak context-memory conflict on counterfactual data, study finds
A new research paper explores how the perceived plausibility of input data affects the faithfulness of large language models (LLMs). The study generated text in multiple languages, including low-resource ones like Czech…
-
Multilingual NLP models detect harmful and verifiable social media posts
Researchers have developed multilingual transformer-based NLP models capable of detecting social media posts that contain verifiable factual claims and harmful content. This study involved dataset collection, pre-proces…
-
New research analyzes politician speech patterns across four Slavic languages
A new research paper details three studies on spoken parliamentary speech across four Slavic languages: Croatian, Czech, Polish, and Serbian. The research analyzed over 6,000 hours of speech from the ParlaSpeech 3.0 cor…
-
Linguistic study visualizes fuzzy nature of parts of speech using word embeddings
Researchers have explored the semantic space of parts of speech using word2vec embeddings and a neural network to reduce dimensionality. This analysis maps thousands of words into a three-dimensional space, revealing pr…
-
Unicode CLDR defines 1-6 plural forms for language localization
The number of plural forms a language requires for software localization can range from one to six, as defined by the Unicode Consortium's Common Locale Data Repository (CLDR). This count is not a linguistic absolute bu…
-
Bengali LLM support lags due to data scarcity, not demand
Despite having a large number of native speakers, Bengali language support in large language models lags significantly behind languages with fewer speakers. This disparity is not due to a lack of demand but rather a sca…
-
AI boom strains hardware supply chains for Czech firms
Czech companies are experiencing significant delays in receiving hardware, with lead times stretching to several months. This shortage is attributed to the immense demand for manufacturing capacity being consumed by the…
-
Horror-simulation game about Czech train conductors released
A new game has been identified as a perfect summer pastime, offering a unique blend of horror and simulation. The game centers around Czech train conductors and features a deep simulation of a fictional subway system, c…
-
Transformers analyze filled pauses in Slavic parliamentary speech
Researchers have utilized transformer-based models to analyze approximately 4,000 hours of parliamentary speech from four Slavic languages: Croatian, Czech, Polish, and Serbian. The study investigated the occurrence and…
-
New corpora reveal human variation in coreference and discourse annotation
Researchers have introduced two new corpora, Hlava Cor and Hlava AD, designed to study human label variation in coreference and discourse relations. Hlava Cor contains 1,024 contexts annotated by three individuals, focu…