Amharic
PulseAugur coverage of Amharic — every cluster mentioning Amharic across labs, papers, and developer communities, ranked by signal.
3 day(s) with sentiment data
-
AI language model costs surge for non-English queries due to tokenization
The cost of using AI language models can vary significantly based on the language of the input, with non-English languages often requiring substantially more tokens for the same query. This is because models are primari…
-
New LoWAgent benchmark reveals performance gaps in AI web agents for low-resource languages
A new benchmark called LoWAgent has been introduced to evaluate the performance of web agents across low-resource languages, addressing a gap in current testing methodologies. The benchmark includes six localized websit…
-
AI agent guards vulnerable to language-based attacks, study finds
A new study published on arXiv reveals that AI agent guards, designed to ensure safety before tool execution, can be vulnerable to language-based attacks. Researchers found that translating user requests into different …
-
New AI routing methods enhance reasoning and efficiency · 4 sources tracked
Researchers have developed T-Router, a parameter-efficient reinforcement learning method that significantly improves reasoning capabilities by selectively reusing computations from pretrained models. This approach, deta…
-
Study finds LLM data audits don't guarantee downstream utility for African NLP
A new study published on arXiv investigates the effectiveness of synthetic data selection methods for low-resource African languages. The research found that common proxies, which assume that data rated highly by an LLM…
-
Neural combinatorial optimization solvers may overstate gains, study finds
A new research paper published on arXiv investigates the effectiveness of test-time budget allocation for neural combinatorial optimization solvers. The study found that while uniform sampling appears to show gains on i…
-
LLM Safety Fails to Transfer Across Low-Resource Languages, Study Finds
A new research paper published on arXiv highlights significant limitations in the cross-lingual safety of large language models (LLMs). The study focused on four African languages—Twi, Hausa, Amharic, and Swahili—and fo…
-
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's impact on linguistic diversity in World Englishes debated in new papers · 3 sources tracked
Three recent academic papers explore the complex relationship between generative AI and linguistic diversity, particularly concerning World Englishes. The first paper discusses how AI tools can both democratize academic…
-
New corpus of US religious radio transcripts released
Researchers have created a large-scale corpus of religious radio broadcast transcripts from webstream recordings in the United States. This dataset, captured over a month in July 2025, includes over 700,000 recordings a…
-
New VEXMLM model boosts performance for African languages
Researchers have developed VEXMLM, a new language model designed to improve performance on low-resource African languages, specifically Amharic and Tigrinya, which use the Ge'ez script. This model addresses issues of hi…
-
New Corpus Aims to Boost African Languages in Science
A new parallel corpus called AfriScience-MT has been developed to address the lack of scientific terminology in six African languages: Amharic, Hausa, Luganda, Northern Sotho, Yorùbá, and isiZulu. This corpus, created b…
-
Hubness hinders multilingual AI retrieval; Amharic needs in-language tuning
Research indicates that cross-lingual retrieval in multilingual embedding models is hindered by "hubness," a geometric pathology in embedding spaces, rather than anisotropy. Studies using models like Gemini, Mistral, an…
-
CURE-Med framework enhances multilingual medical reasoning in LLMs
Researchers have developed CURE-Med, a novel framework using curriculum-informed reinforcement learning to enhance multilingual medical reasoning in large language models. This approach integrates code-switching-aware s…