Prosecution Insights
Prosecution Insights
Tech Center 4100 • Art Units: 2122 2126 2128 4100
This examiner grants 50% of resolved cases
Molly Clarke Sippel examines applications in Art Unit 2122 under Tech Center 4100. With 28 resolved cases, the examiner has an allowance rate of 50.0%. This indicates that 50.0% of the applications result in a grant. The interview lift of 25.7% suggests that direct engagement is a valuable tool for advancing prosecution. The median time to allowance is 1398 days.
Practitioners should factor the 50.0% allowance rate and the 1398-day timeline into their strategy for Art Unit 2122. The 25.7% interview lift provides a significant boost for applicants. Across 28 resolved cases, the examiner has shown that verbal advocacy is effective in Tech Center 4100. Strategy should focus on using the 25.7% lift to improve the baseline grant probability. The 1398-day median pendency reflects the expected duration for successful outcomes in this docket.
| App # | Title | Status | Assignee |
|---|---|---|---|
| 18639937 | SYSTEM AND METHOD FOR HARDWARE-AWARE PRUNING OF CONFORMER NETWORKS | Non-Final OA | Samsung Electronics Co., Ltd. |
| 18518753 | FEDERATED LEARNING METHOD, APPARATUS, AND SYSTEM | Non-Final OA | HUAWEI TECHNOLOGIES CO., LTD. |
| 18175936 | Deep Learning Training Method for Computing Device and Apparatus | Non-Final OA | Huawei Technologies Co., Ltd. |
| 17520197 | DEVICE, METHOD AND SYSTEM FOR REGULARIZATION OF A BINARY NEURAL NETWORK | Final Rejection | HUAWEI TECHNOLOGIES CO., LTD. |
| 18217006 | FEDERATED LEARNING WITH MODEL DIVERSITY AND BACKUP IN CASE OF DISCONNECTED CLIENT | Final Rejection | Robert Bosch GmbH |
| 18216982 | FEDERATED LEARNING WITH MODEL DIVERSITY | Final Rejection | Robert Bosch GmbH |
| 18448896 | SYSTEMS AND METHODS FOR TRAINING REINFORCEMENT LEARNING MODELS USING UNSUPERVISED MODELS | Final Rejection | Capital One Services, LLC |
| 18635391 | System And Method For Dynamic Hyperparameter Optimization For Large Language Models Using (Few-Shot) Reinforcement Learning | Non-Final OA | Oracle International Corporation |
| 18009425 | GROUPING NODES IN A SYSTEM | Non-Final OA | Telefonaktiebolaget LM Ericsson (publ) |
| 18079123 | PROVIDING TRAINED REINFORCEMENT LEARNING SYSTEMS | Final Rejection | MASSACHUSETTS INSTITUTE OF TECHNOLOGY |
| 18597648 | APPARATUS AND METHOD FOR DATA LEARNING SYSTEM BASED ON A SECURED MODELING EXCHANGE | Non-Final OA | NTT DOCOMO, INC. |
| 18176037 | TRANSFORMER NETWORK WITH NORMALIZATION INCLUDING SCALING PARAMETER | Non-Final OA | Microsoft Technology Licensing, LLC |
| 18488246 | MULTI-TASK OFFLINE REINFORCEMENT LEARNING MODEL BASED ON SKILL REGULARIZED TASK DECOMPOSITION AND MULTI-TASK OFFLINE REINFORCEMENT LEARNING METHOD USING THE SAME | Non-Final OA | RESEARCH & BUSINESS FOUNDATION SUNGKYUNKWAN UNIVERSITY |
| 17972135 | TOPOLOGY-AUGMENTED SYSTEM FOR AI-MODEL MISMATCH | Final Rejection | Raytheon Company |
| 18180380 | GRAPH BASED PREDICTIVE INFERENCES FOR DOMAIN TAXONOMY | Non-Final OA | Optum, Inc. |
| 18639686 | BLACK-BOX OPTIMIZATION USING NEURAL NETWORKS | Non-Final OA | DeepMind Technologies Limited |
IP Author analyzes examiner patterns and generates tailored response strategies with the highest chance of allowance.
Build Your Strategy