This inventor holds 5 USPTO granted patents. Top assignee: Intuit, Inc.. Active years: 2026.
Company Filing History:
Years Active: 2026
Title: Amir Bialer: Innovator in Language Model Technologies
Introduction
Amir Bialer is a prominent inventor based in Tel Aviv, Israel. He has made significant contributions to the field of language model technologies, holding a total of 5 patents. His work focuses on enhancing the accuracy and reliability of generative language models.
Latest Patents
One of Amir's latest patents is titled "Unsupervised detection of hallucinations in language model outputs." This innovative method provides a way to detect hallucinations in generative language model outputs without supervision. The process involves receiving a language model output that includes tokens and their corresponding confidence scores. It identifies low-confidence tokens whose scores fall below a specified threshold and determines unsupervised attributes of these tokens. An anomaly score is calculated based on these attributes, and if the score meets or exceeds a predetermined threshold, the system predicts the presence of at least one hallucination in the output.
Another notable patent is "Closed-loop speculative decoding with real-time prompt patching for language model output generation." This method involves generating a token sequence in response to an input prompt using a draft model. Concurrent operations classify the token sequence and verify its semantic accuracy against an accuracy threshold. If the token sequence does not meet the accuracy requirements, it is regenerated, ensuring high-quality output.
Career Highlights
Amir Bialer is currently employed at Intuit, Inc., where he continues to develop innovative solutions in the realm of language processing. His expertise in this area has positioned him as a key player in advancing the capabilities of generative language models.
Collaborations
Amir has collaborated with notable colleagues such as Shai Ardazi and Ofir Ben Shoham. Their combined efforts contribute to the ongoing research and development in language model technologies.
Conclusion
Amir Bialer's work exemplifies the innovative spirit in the field of language models. His patents reflect a commitment to improving the accuracy and reliability of generative outputs. Through his contributions, he continues to shape the future of language processing technologies.