Patents by Inventor Weiran YAO

Weiran YAO has filed for patents to protect the following inventions. This listing includes patent applications that are pending as well as patents that have already been granted by the United States Patent and Trademark Office (USPTO).

  • Publication number: 20250259054
    Abstract: Embodiments described herein provide a unified LLM training pipeline that hands the diversity of various data structures and formats involving LLMs agent trajectories. These pipelines are specifically designed to transform incoming data into a standardized representation, ensuring compatibility across varied formats. Furthermore, the data collection undergoes a filtering process to ensure high-quality trajectories, adding an additional layer of refinement to the dataset. In this way, the training pipeline not only unifies trajectories across environments but also enhances the overall quality and reliability of the collected data for LLM training.
    Type: Application
    Filed: May 8, 2024
    Publication date: August 14, 2025
    Inventors: Jianguo Zhang, Tian Lan, Rithesh Murthy, Zhiwei Liu, Weiran Yao, Juntao Tan, Shelby Heinecke, Yihao Feng, Huan Wang, Juan Carlos Niebles, Silvio Savarese, Caiming Xiong
  • Publication number: 20250139411
    Abstract: Embodiments described herein provide a large language model (LLM) based AI agent that adopts Monte-Carlo Tree Search (MCTS) to execute a task. The LLM is prompted with a task description and it responds with its first attempted list of actions. Based on the success or failure of the first attempt, the LLM is prompted with an updated prompt which includes feedback from the first attempt based on a determined reward. The prompt may include a relative “score” for each action taken at each step. A numeric score may be mapped to a set of pre-defined text labels, such as “high” or “low” value putting the score in a form more suited for an LLM prompt. In this way, the LLM is iteratively given prompts which are updated with the scores from each action taken at each previous iterations so that it traverses different paths on the tree in each iteration.
    Type: Application
    Filed: October 31, 2023
    Publication date: May 1, 2025
    Inventors: Rithesh Murthy, Shelby Heinecke, Juan Carlos Niebles Duque, Zhiwei Liu, Le Xue, Weiran Yao, Yihao Feng, Zeyuan Chen, Akash Gokul, Devansh Arpit, Ran Xu, Lik Mui, Huan Wang, Caiming Xiong, Silvio Savarese
  • Publication number: 20250124233
    Abstract: Systems and methods for editing a large language model are provided. The large language model generates a sequence of tokens, a first probability of a pre-edit output based on the sequence of tokens, and a second probability of a target output based on the sequence of tokens. A loss function is provided based on the first probability and the second probability. A plurality of gradients of the large language model with respect to the loss function is computed. An edit location of the large language model is determined based on the plurality of gradients. The large language model is edited by editing weights at the edit location of the large language model, such that the updated large language model generates the target output for an input including the sequence of words.
    Type: Application
    Filed: January 31, 2024
    Publication date: April 17, 2025
    Inventors: Itai Izhak Feigenbaum, Devansh Arpit, Shelby Heinecke, Juan Carlos Niebles Duque, Weiran Yao, Huan Wang, Caiming Xiong, Silvio Savarese
  • Publication number: 20250053793
    Abstract: Embodiments described herein provide a method of predicting an action by a plurality of language model augmented agents (LAAs). In at least one embodiment, a controller receives a task instruction to be performed using an environment. The controller receives an observation of a first state from the environment. The controller selects a LAA from the plurality of LAAs based on the task instruction and the observation. The controller obtains an output from the selected LAA generated using an input combining the task instruction, the observation, and an LAA-specific prompt template. The controller determines the action based on the output. The controller causes the action to be performed on the environment thereby causing the first state of the environment to change to a second state.
    Type: Application
    Filed: October 25, 2023
    Publication date: February 13, 2025
    Inventors: Zhiwei Liu, Weiran Yao, Jianguo Zhang, Le Xue, Shelby Heinecke, Rithesh Murthy, Yihao Feng, Zeyuan Chen, Juan Carlos Niebles Duque, Devansh Arpit, Ran Xu, Lik Mui, Huan Wang, Caiming Xiong, Silvio Savarese
  • Publication number: 20250045567
    Abstract: Embodiments described herein provide for optimizing a language model (LM) agent. In at least one embodiment, and LM agent comprises an “actor” LM and a “retrospective LM which provides reflections on attempts by the actor LM. The reflections are used to update subsequent prompts to the actor LM. Optimizing the LM agent comprises fine-tuning parameters of the retrospective LM while keeping parameters of the actor LM frozen. A gradient may be determined by a change in reward from the environment based on actions taken by the actor LM with and without a reflection of the retrospective LM. Using this gradient, parameters of the retrospective LM may be updated via backpropagation.
    Type: Application
    Filed: October 31, 2023
    Publication date: February 6, 2025
    Inventors: Weiran Yao, Shelby Heinecke, Juan Carlos Niebles Duque, Zhiwei Liu, Yihao Feng, Le Xue, Rithesh Murthy, Zeyuan Chen, Jianguo Zhang, Devansh Arpit, Ran Xu, Lik Mui, Huan Wang, Caiming Xiong, Silvio Savarese
  • Patent number: 12027858
    Abstract: A computer implemented method for controlling a load aggregator for a grid includes receiving a predicted power demand over a horizon of time steps associated with one of at least two buildings, aggregating the predicted power demand at each time step to obtain an aggregate power demand, applying a learnable convolutional filter on the aggregate power demand to obtain a target load, computing a difference between the predicted power demand of the one building with the target load to obtain a power shift associated with the one building over the horizon of time steps, apportioning the power shift according to a learnable weighted vector to obtain an apportioned power shift, optimizing the learnable weighted vector and the learnable convolutional filter via an evolutionary strategy based update to obtain an optimized apportioned power shift, and transmitting the optimized apportioned power shift to a building level controller associated with the one building.
    Type: Grant
    Filed: July 1, 2021
    Date of Patent: July 2, 2024
    Assignees: Robert Bosch GmbH, Carnegie Mellon University
    Inventors: Jonathan Francis, Bingqing Chen, Weiran Yao
  • Publication number: 20230025215
    Abstract: A computer implemented method for controlling a load aggregator for a grid includes receiving a predicted power demand over a horizon of time steps associated with one of at least two buildings, aggregating the predicted power demand at each time step to obtain an aggregate power demand, applying a learnable convolutional filter on the aggregate power demand to obtain a target load, computing a difference between the predicted power demand of the one building with the target load to obtain a power shift associated with the one building over the horizon of time steps, apportioning the power shift according to a learnable weighted vector to obtain an apportioned power shift, optimizing the learnable weighted vector and the learnable convolutional filter via an evolutionary strategy based update to obtain an optimized apportioned power shift, and transmitting the optimized apportioned power shift to a building level controller associated with the one building.
    Type: Application
    Filed: July 1, 2021
    Publication date: January 26, 2023
    Inventors: Jonathan FRANCIS, Bingqing CHEN, Weiran YAO