The abundance of web-scale textual data available has been a major factor in the development of generative language models, such as those pretrained as multi-purpose foundation models and tailored for particular Natural Language Processing (NLP) tasks. These models...

Mathematical reasoning is vital for problem-solving and decision-making, particularly in large language models (LLMs). Evaluating LLMs’ mathematical reasoning usually focuses on the final result rather than the reasoning process intricacies. Current methodologies,...

The field of artificial intelligence is rapidly advancing, and there have been significant improvements in text-to-speech (TTS) technology. Parler-TTS is a new open-source inference and training library that has been designed to encourage innovation in high-quality...

Language models often need more exposure to fruitful mistakes during training, hindering their ability to anticipate consequences beyond the next token. LMs must improve their capacity for complex decision-making, planning, and reasoning. Transformer-based models...

Imagine an AI system that can recognize any object, comprehend any text, and generate realistic images without being explicitly trained on those concepts. This is the enticing promise of “zero-shot” capabilities in AI. But how close are we to realizing this vision?...

Speech synthesis has greatly progressed in technological advancements, reflecting the human quest for machines that speak like us. As we stride into an era where interactions with digital assistants and conversational agents become commonplace, the demand for speech...