Bangla
PulseAugur coverage of Bangla — every cluster mentioning Bangla across labs, papers, and developer communities, ranked by signal.
9 day(s) with sentiment data
-
Computational linguistics studies analyze lexical transmission in religious texts · 4 sources tracked
Two new computational linguistics studies analyze lexical transmission and stylistic features within religious texts. The first paper examines Bengali and Sanskrit devotional literature from the 8th to 19th centuries, u…
-
Olfactory-inspired architecture boosts low-resource NER
Researchers have developed a novel, biologically-inspired olfactory architecture to improve Named Entity Recognition (NER) in low-resource languages. This architecture, termed a receptor-glomerular bottleneck, is integr…
-
AI models struggle with multilingual mental health data generation
A new research paper explores the limitations of using persona-based localization to create multilingual mental health datasets. The study found that simply adjusting nationality and language parameters in synthetic per…
-
New dataset and model tackle hate speech in Bengali memes
Researchers have introduced Bn-HIB, a new dataset designed to detect hate and inflammatory content within Bengali internet memes. This dataset, containing 3,247 manually annotated memes, is notable for being the first t…
-
Compact Bangla LLM Outperforms Larger Models with Efficient Design
Researchers have developed a new compact language model, bangla-smollm-135m, specifically designed for the Bangla language. This 135-million parameter model achieves competitive performance against larger models by empl…
-
Bangla language grading system uses fine-tuned lightweight LLM
Researchers have developed a new system for grading written answers in Bangla, a low-resource language, by fine-tuning a lightweight language model. This system prioritizes semantic correctness over exact wording to pro…
-
LLMs struggle with cultural translation in math problems
A new study analyzed how large language models like Claude Opus 4, GPT-4.1, and Gemini 2.5 Pro translate math word problems across various languages and cultures. The research found that while models often agree on the …
-
New dataset boosts personalized book recommendations for Bangla literature
Researchers have introduced RokomariBG, a new large-scale dataset designed to advance personalized book recommendations within Bangla literature. This heterogeneous book graph dataset includes over 127,000 books and rel…
-
New benchmark probes LLM sycophancy in Bengali conversations
Researchers have developed BenSyc, a new benchmark designed to evaluate how large language models exhibit sycophancy within Bengali social conversations. The benchmark, built from Reddit data, categorizes responses into…
-
LLMs struggle with consumer device repair, GPT-5.4 leads
A new benchmark evaluates large language models on their ability to answer real-world consumer device repair questions. The study found that while LLMs can offer some assistance, they are unreliable for high-risk tasks,…
-
Bengali Braille to Text Translator Achieves 97.25% Accuracy
Researchers have developed a geometric approach to translate Bengali Braille characters into text, addressing a gap in available software for visually impaired individuals learning the language. The system processes ima…
-
New Benchmark Evaluates Multilingual VLMs on Bengali Culture and Dialects
Researchers have developed BanglaVerse, a new benchmark designed to evaluate the cultural understanding of multilingual vision-language models (VLMs) within the context of Bengali culture. This benchmark, comprising 1,1…
-
LLMs and audio models show varied success in zero-shot Parkinson's detection
Researchers have compared two methods for using large language and audio models to detect Parkinson's disease from speech without prior training. The study found that performance varied depending on whether the models p…
-
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…
-
New BLADE dataset improves honorifics in multilingual Bangla LLMs
Researchers have developed a new dataset and benchmarking framework called BLADE to address honorific failures in multilingual Bangla text generation. This dataset comprises over 4,000 curated interaction pairs designed…
-
LLMs improve multilingual speech correction by tuning for fluency
Researchers have developed a new method for correcting disfluencies in multilingual speech transcripts using large language models (LLMs). The pipeline first identifies disfluent tokens and then uses these signals to fi…
-
Diffusion augmentation boosts Bangla character recognition accuracy
Researchers have developed a confidence-guided diffusion augmentation method to improve the recognition of handwritten Bangla compound characters. This approach uses diffusion models to generate high-quality synthetic c…
-
New defense framework tackles multilingual prompt injection attacks
Researchers have developed MIPIAD, a defense framework to combat indirect prompt injection attacks in multilingual large language model systems. The framework combines a Qwen2.5-1.5B model fine-tuned with LoRA, TF-IDF l…
-
Bengali AI models show identity biases despite similar data, study finds
A new paper investigates biases in sentiment analysis models for the Bengali language, a low-resource context. Researchers audited models like mBERT and BanglaBERT, fine-tuned on Bengali sentiment analysis datasets, and…
-
Llama-3.2-3B model achieves 92% accuracy in parsing blood donation requests
Researchers have developed the Cognitive Blood Request System (CBRS), a framework designed to efficiently filter and parse urgent blood donation requests from social media streams. This system utilizes a novel bilingual…