Whisper
PulseAugur coverage of Whisper — every cluster mentioning Whisper across labs, papers, and developer communities, ranked by signal.
- 2026-06-09 research_milestone A study on fine-tuning OpenAI's Whisper for Swiss German ASR revealed improved performance and identified benchmark contamination issues. source
- 2026-05-12 research_milestone A new semi-supervised framework for speech confidence detection was proposed, achieving a Macro-F1 score of 0.751. source
19 day(s) with sentiment data
-
New Hungarian ASR Corpus Doubles Training Data, Improves Accuracy
Researchers have introduced BEA-Dialogue+, an expanded corpus for Hungarian conversational automatic speech recognition (ASR). This new dataset increases the available training data to 200 hours, relaxing split criteria…
-
Research Support Hub offers AI and data analysis workshops
The Research Support Hub is offering a series of workshops in June and July, covering topics such as QGIS, AI tools for research, and transcription with Whisper and LLMs. Other sessions will focus on cloud computing, AI…
-
Nvidia RTX 1650 powers local Whisper AI transcription
An individual repurposed an unused Nvidia RTX 1650 graphics card by installing it in a server to create a local instance of OpenAI's Whisper speech-to-text service. This setup allows for private, on-premises transcripti…
-
New text-only method adapts speech recognition models
Researchers have developed WhisTLE, a novel method for adapting pre-trained automatic speech recognition (ASR) models using only text data. This technique employs a variational autoencoder to model encoder outputs and f…
-
Voice-channel tool enables hands-free control of multiple Claude Code Agents
A developer has created a new open-source tool called voice-channel that allows users to control multiple Claude Code Agents using hands-free voice commands. The system, designed for local network use, routes spoken com…
-
AI integration market matures, focusing on depth over new launches
The MCP ecosystem experienced a quiet week with no new server launches, indicating a maturing market where developers are prioritizing deeper integrations over novelty. Usage is consolidating around established, free se…
-
AI participation tools show bias against non-Western names and accents
AI tools designed to track meeting participation and contribution are showing bias against non-Western names and accents. These systems, used by companies like Amazon and Meta, are trained on data that underrepresents c…
-
New Windows app SEELS enables local LLM training via user corrections
A new Windows desktop application called SEELS has been released, designed for running local Large Language Models (LLMs). Its core feature allows users to correct model responses and use these corrections to train cust…
-
Thinking Machines unveils real-time interaction models with 200ms processing
Thinking Machines has unveiled a new class of "interaction models" designed for real-time conversational AI. These models process audio, video, and text in rapid 200-millisecond intervals, eliminating the need for separ…
-
Convex optimization framework boosts accent-robust language detection
Researchers have developed a new convex optimization framework called Convex Language Detection (CLD) to improve language identification in speech recognition systems, particularly for low-resource accents and dialects.…
-
Developer builds Hindi voice-to-form app for health workers
A developer built Sakhi, a Hindi voice-to-form application for India's community health workers, in six weeks. The system addresses challenges with unreliable cloud speech-to-text and intermittent connectivity in rural …
-
OpenAI launches GPT Realtime 2; Anthropic expands Claude for Legal
OpenAI has launched new voice intelligence features, including GPT Realtime 2 powered by GPT-5, offering real-time translation and transcription with an emphasis on reduced latency and larger context windows. Anthropic …
-
AI tools like LLMs can now be run on personal hardware
A Golem.de article explores how to run large language models (LLMs) and other AI tools like Whisper locally on personal hardware. It discusses the increasing feasibility of self-hosting these technologies, moving away f…
-
Vector RAG vs. LLM Wiki: Study reveals trade-offs in research synthesis
A new research paper compares Vector Retrieval-Augmented Generation (RAG) against an LLM-compiled wiki for answering questions over a small corpus of 24 research papers. While the wiki excelled at synthesizing informati…
-
Developer tool extracts code from videos using local AI
A developer has created a local tool called videocode that extracts runnable code from video tutorials. The tool utilizes scene detection, audio transcription via Whisper, and vision models like LLaVA and Llama3.2-visio…
-
New benchmark tackles ASR bias in Indic languages
Researchers have developed Vividh-ASR, a new benchmark designed to evaluate automatic speech recognition (ASR) models for Indic languages, specifically Hindi and Malayalam. This benchmark categorizes audio into four tie…
-
CognitiveBotics builds personalized AI content engine for autistic children
CognitiveBotics has developed a personalized content engine for children with autism, addressing the challenge of high individual variability in learning preferences. Their Modalities Engine renders learning objectives …
-
New framework improves speech confidence detection using Whisper
Researchers have developed a new semi-supervised framework for detecting speaker confidence in speech, addressing the challenge of limited labeled data. This approach combines deep semantic embeddings from OpenAI's Whis…
-
Developer releases llmclean library to clean LLM output
A developer has released version 0.2.0 of llmclean, a Python library designed to clean and normalize output from large language models. The library addresses common issues such as removing markdown fences, repairing mal…
-
MCP Ecosystem Matures: Official Integrations Dominate Developer Attention
The MCP ecosystem is maturing, with a focus shifting from adding new servers to refining existing integrations. Official integrations from major platforms like GitHub, OpenAI, and Figma are dominating developer attentio…