Prosecution Insights
Prosecution Insights
Technology area: Computing & Software
52 pending office actions • 19 art units • 50 examiners • 9 of 52 (17%) have an AI response strategy ready • 19 patents granted in the last 365 days
DeepMind Technologies Limited maintains a robust portfolio with 47 pending office actions in the Computing & Software technology area. These actions are distributed across 45 distinct examiners, suggesting a broad exposure to different examination styles and individual preferences. The portfolio is further diversified across 19 distinct art units, reflecting the wide-ranging technical applications of their software innovations and artificial intelligence research.
The busiest examiner, TRAN, TAN H, is currently managing 2 pending office actions. This low concentration of cases per examiner, with nearly as many distinct examiners as there are pending actions, indicates that the company's prosecution is highly decentralized. Managing the various procedural nuances and technical standards across so many different sections of the USPTO requires a sophisticated strategy.
The high volume of pending actions highlights the active and expansive nature of their current patenting efforts. With 45 distinct examiners involved, the company must prepare for a wide variety of feedback and potential rejections. The distribution across 19 distinct art units means that their legal team interacts with many different administrative groups, each with its own internal culture and precedent, suggesting the company prioritizes a broad intellectual property footprint.
Based on the USPTO statutory response window for each pending office action. 6 of the docket's apps have a known mailing date; the rest are excluded from the tile counts.
Every pending office action with a known statutory deadline, placed on a days-until-due axis. Dots left of Today are overdue; the further right, the more runway. Cases that share a deadline window stack vertically. 6 of the docket's apps have a known mailing date.
Difficulty is derived from the rejection statutes on the most recent pending office action. §101-driven and multi-statute cases are graded Hard; §112-only and obviousness-type double-patenting cases are graded Easy; everything else is Medium. "Unknown" means we have not yet parsed a statute for that office action.
| Bucket | Cases |
|---|---|
| §101 only | 2 (4%) |
| §101 + other | 24 (46%) |
| §103 only | 14 (27%) |
| §112 only | 2 (4%) |
| Double-patenting only | 2 (4%) |
| Multi-statute (no §101) | 8 (15%) |
How the docket's pending cases split across USPTO tech-center bands.
Manual office-action response work runs about 10 hours per case. The time-saved bands below show what IP Author's prosecution pipeline typically delivers — a conservative 20% on the low end, 35% in the middle, 50% on the high end.
| Examiner | Apps on this docket | Allow rate | Interview lift |
|---|---|---|---|
| TRAN, TAN H | 2 | 60.9% | +32.6% |
| FONSECA LOPEZ, FRANCINI ALVARENGA | 2 | 29.6% | +37.1% |
| BUKSA, CHRISTOPHER ALLEN | 1 | 73.6% | +20.1% |
| REFAI, RAMSEY | 1 | 50.7% | +12.3% |
| COLUCCI, MICHAEL C | 1 | 75.8% | +15.3% |
| PHAM, ANNIE | 1 | 87.5% | +14.3% |
| LIU, XIAO | 1 | 87.7% | +12.0% |
| BEZUAYEHU, SOLOMON G | 1 | 75.7% | +29.9% |
| ELISCA, PIERRE E | 1 | 90.2% | +6.4% |
| KASSIM, HAFIZ A | 1 | 44.7% | +53.8% |
Cases in front of an examiner with an allow rate of 80%+ where the difficulty is Easy or Medium. The top 2 ordered by deadline are shown.
| App # | Title | Examiner | Due in |
|---|---|---|---|
| 18951101 | GENERATING TRAINING DATA USING A GENERATIVE NEURAL NETWORK | PHAM, ANNIE | — |
| 18919257 | RATING TASKS AND POLICIES USING CONDITIONAL PROBABILITY DISTRIBUTIONS DERIVED FROM EQUILIBRIUM-BASED SOLUTIONS OF GAMES | ELISCA, PIERRE E | — |
Multi-statute / §101-driven matters, or cases in front of an examiner with an allow rate under 30%. The top 8 ordered by deadline are shown.
| App # | Title | Examiner | Due in |
|---|---|---|---|
| 18277729 | NOWCASTING USING GENERATIVE NEURAL NETWORKS | ISHIZUKA, YOSHIHISA | 38d overdue |
| 18127551 | ALLOCATING COMPUTING RESOURCES BETWEEN MODEL SIZE AND TRAINING DATA DURING TRAINING OF A MACHINE LEARNING MODEL | TAN, DAVID H | 38d overdue |
| 17959232 | NEURAL NETWORKS WITH TRANSFORMED ACTIVATION FUNCTION LAYERS | HADDAD, MAJD MAHER | 31d overdue |
| 19111985 | TRAINING POLICY NEURAL NETWORKS IN SIMULATION USING SCENE SYNTHESIS MACHINE LEARNING MODELS | BUKSA, CHRISTOPHER ALLEN | — |
| 18878496 | SIMULATING INDUSTRIAL FACILITIES FOR CONTROL | REFAI, RAMSEY | — |
| 18986673 | ADAPTIVE VISUAL SPEECH RECOGNITION | COLUCCI, MICHAEL C | — |
| 18945375 | UNSUPERVISED LEARNING OF OBJECT KEYPOINT LOCATIONS IN IMAGES THROUGH TEMPORAL TRANSPORT OR SPATIO-TEMPORAL TRANSPORT | LIU, XIAO | — |
| 18843608 | AGENT CONTROL THROUGH CULTURAL TRANSMISSION | KASSIM, HAFIZ A | — |
Cases in front of an examiner whose interview lift is 10 percentage points or more — i.e. interviewed cases historically resolve more favorably than non-interviewed ones. The top 8 ordered by deadline are shown.
| App # | Title | Examiner | Due in |
|---|---|---|---|
| 19111985 | TRAINING POLICY NEURAL NETWORKS IN SIMULATION USING SCENE SYNTHESIS MACHINE LEARNING MODELS | BUKSA, CHRISTOPHER ALLEN | — |
| 18878496 | SIMULATING INDUSTRIAL FACILITIES FOR CONTROL | REFAI, RAMSEY | — |
| 18986673 | ADAPTIVE VISUAL SPEECH RECOGNITION | COLUCCI, MICHAEL C | — |
| 18951101 | GENERATING TRAINING DATA USING A GENERATIVE NEURAL NETWORK | PHAM, ANNIE | — |
| 18945375 | UNSUPERVISED LEARNING OF OBJECT KEYPOINT LOCATIONS IN IMAGES THROUGH TEMPORAL TRANSPORT OR SPATIO-TEMPORAL TRANSPORT | LIU, XIAO | — |
| 18935365 | TRAINING IMAGE REPRESENTATION NEURAL NETWORKS USING CROSS-MODAL INTERFACES | BEZUAYEHU, SOLOMON G | — |
| 18843608 | AGENT CONTROL THROUGH CULTURAL TRANSMISSION | KASSIM, HAFIZ A | — |
| 18668080 | CONTROLLING AGENTS BY TRANSFERRING SUCCESSOR FEATURES TO NEW TASKS | TRAN, TAN H | — |
| Art Unit | Apps |
|---|---|
| 2143 | 3 |
| 2124 | 2 |
| 2126 | 2 |
| 2146 | 2 |
| 2125 | 2 |
| 2145 | 2 |
| 2123 | 2 |
| 3658 | 1 |
| 2662 | 1 |
| 2677 | 1 |
| App # | Title | Examiner | Art Unit | Statutes | Status | Due in | AI | Filed |
|---|---|---|---|---|---|---|---|---|
| 19111985 | TRAINING POLICY NEURAL NETWORKS IN SIMULATION USING SCENE SYNTHESIS MACHINE LEARNING MODELS | BUKSA, CHRISTOPHER ALLEN | 3658 | §102§103 | Non-Final OA | — | Pending | Mar 14, 2025 |
| 18878496 | SIMULATING INDUSTRIAL FACILITIES FOR CONTROL | REFAI, RAMSEY | — | §101§102 | Non-Final OA | — | Pending | Dec 23, 2024 |
| 18986673 | ADAPTIVE VISUAL SPEECH RECOGNITION | COLUCCI, MICHAEL C | — | §102§103 | Non-Final OA | — | Pending | Dec 18, 2024 |
| 18951101 | GENERATING TRAINING DATA USING A GENERATIVE NEURAL NETWORK | PHAM, ANNIE | 2662 | §103 | Non-Final OA | — | Pending | Nov 18, 2024 |
| 18945375 | UNSUPERVISED LEARNING OF OBJECT KEYPOINT LOCATIONS IN IMAGES THROUGH TEMPORAL TRANSPORT OR SPATIO-TEMPORAL TRANSPORT | LIU, XIAO | — | §102§103Other | Non-Final OA | — | Pending | Nov 12, 2024 |
| 18935365 | TRAINING IMAGE REPRESENTATION NEURAL NETWORKS USING CROSS-MODAL INTERFACES | BEZUAYEHU, SOLOMON G | — | §103 | Non-Final OA | — | Pending | Nov 01, 2024 |
| 18919257 | RATING TASKS AND POLICIES USING CONDITIONAL PROBABILITY DISTRIBUTIONS DERIVED FROM EQUILIBRIUM-BASED SOLUTIONS OF GAMES | ELISCA, PIERRE E | — | DP | Non-Final OA | — | Pending | Oct 17, 2024 |
| 18843608 | AGENT CONTROL THROUGH CULTURAL TRANSMISSION | KASSIM, HAFIZ A | — | §101§103 | Non-Final OA | — | Pending | Sep 03, 2024 |
| 18794786 | USING VISUAL LANGUAGE MODELS TO DETERMINE LOCATIONS OF IMAGE ELEMENTS WITHIN GRAPHICAL IMAGES | PATEL, JAYESH A | 2677 | §101 | Non-Final OA | — | Pending | Aug 05, 2024 |
| 18834208 | CONTROLLING REINFORCEMENT LEARNING AGENTS USING GEOMETRIC POLICY COMPOSITION | JANSEN II, MICHAEL J | — | §101§102§103 | Non-Final OA | — | Pending | Jul 29, 2024 |
| 18754726 | DISTRIBUTIONAL REINFORCEMENT LEARNING | HUANG, YAO D | 2124 | §103Other | Final Rejection | — | Pending | Jun 26, 2024 |
| 18668080 | CONTROLLING AGENTS BY TRANSFERRING SUCCESSOR FEATURES TO NEW TASKS | TRAN, TAN H | — | §103 | Non-Final OA | — | Pending | May 17, 2024 |
| 18665702 | VERIFICATION OF AGENT OUTPUT THROUGH ADVERSARIAL DEBATE | VO, STEVEN | — | §101§103 | Non-Final OA | — | Pending | May 16, 2024 |
| 18666682 | GENERATING PREDICTIONS FOR NON-STATIONARY DATA USING DISTRIBUTIONS OVER OUTPUT HEAD WEIGHTS | RUDY, ANDREW J | — | §112 | Non-Final OA | — | Pending | May 16, 2024 |
| 18661188 | Dynamic Controlled Decoding | TRAN, QUOC A | — | §112 | Non-Final OA | — | Pending | May 10, 2024 |
| 18639686 | BLACK-BOX OPTIMIZATION USING NEURAL NETWORKS | SIPPEL, MOLLY CLARKE | 2126 | §101Other | Non-Final OA | — | Pending | Apr 18, 2024 |
| 18636971 | TRAINING A POPULATION OF ADVERSARIAL NEURAL NETWORKS TO IMPROVE A BASE NEURAL NETWORK | HARPER, ELIYAH STONE | — | §102§103 | Non-Final OA | — | Pending | Apr 16, 2024 |
| 18698218 | DEMONSTRATION-DRIVEN REINFORCEMENT LEARNING | MRABI, HASSAN | — | §103 | Non-Final OA | — | Pending | Apr 03, 2024 |
| 18623952 | GENERATING DISCRETE LATENT REPRESENTATIONS OF INPUT DATA ITEMS | WU, NICHOLAS S | — | DP | Non-Final OA | — | Pending | Apr 01, 2024 |
| 18424437 | METHODS AND SYSTEMS FOR CONSTRAINED REINFORCEMENT LEARNING | LAU, KAITLYN RENEE | — | §101§102§103§112 | Non-Final OA | — | Pending | Jan 26, 2024 |
| 18424687 | GENERATING ENVIRONMENT MODELS USING IN-CONTEXT ADAPTATION AND EXPLORATION | ACOSTA, RILEY SULLIVAN | — | §101§103§112 | Non-Final OA | — | Pending | Jan 26, 2024 |
| 18422620 | NEURAL POPULATION LEARNING | MAMILLAPALLI, PALLAVI | — | §101§103§112 | Non-Final OA | — | Pending | Jan 25, 2024 |
| 18423239 | TRAINING A NEURAL NETWORK TO PERFORM AN ALGORITHMIC TASK USING A SELF-SUPERVISED LOSS | HICKS, AUSTIN JAMES | — | §103§112 | Non-Final OA | — | Pending | Jan 25, 2024 |
| 18406995 | CONTROLLING AGENTS USING AMORTIZED Q LEARNING | CHEN, ALAN S | — | §103§112 | Non-Final OA | — | Pending | Jan 08, 2024 |
| 18527211 | HIERARCHICAL TEXT GENERATION USING LANGUAGE MODEL NEURAL NETWORKS | TRACY JR., EDWARD | 2656 | §103 | Non-Final OA | — | Pending | Dec 01, 2023 |
| 18285519 | LEARNING DIVERSE SKILLS FOR TASKS USING SEQUENTIAL LATENT VARIABLES FOR ENVIRONMENT DYNAMICS | FIGUEROA, KEVIN W | 2124 | §101§103 | Non-Final OA | — | Pending | Oct 04, 2023 |
| 18374447 | DISCRETE TOKEN PROCESSING USING DIFFUSION MODELS | ROY, SANCHITA | 2146 | §101§103§112 | Non-Final OA | — | Pending | Sep 28, 2023 |
| 18283131 | TRAINING GRAPH NEURAL NETWORKS USING A DE-NOISING OBJECTIVE | TRAN, TAN H | 2141 | §103 | Final Rejection | — | Pending | Sep 20, 2023 |
| 18278473 | CONTINUAL LEARNING NEURAL NETWORK SYSTEM TRAINING FOR CLASSIFICATION TYPE TASKS | LAHAM BAUZO, ALVARO SALIM | 2146 | §103 | Final Rejection | — | AI Ready | Aug 23, 2023 |
| 18277729 | NOWCASTING USING GENERATIVE NEURAL NETWORKS | ISHIZUKA, YOSHIHISA | 2857 | §101§103§112 | Final Rejection | 38d overdue | AI Ready | Aug 17, 2023 |
| 18275722 | IMITATION LEARNING BASED ON PREDICTION OF OUTCOMES | KARTHOLY, REJI P | 2143 | §101§103 | Non-Final OA | — | Pending | Aug 03, 2023 |
| 18230056 | CONTROLLING AGENTS USING AUXILIARY PREDICTION NEURAL NETWORKS THAT GENERATE STATE VALUE ESTIMATES | CHEN, KUANG FU | 2143 | §103 | Final Rejection | 50d | AI Ready | Aug 03, 2023 |
| 18275332 | RENDERING NEW IMAGES OF SCENES USING GEOMETRY-AWARE NEURAL NETWORKS CONDITIONED ON LATENT VARIABLES | AHN, CHRISTINE YERA | 2615 | §103 | Final Rejection | 39d | AI Ready | Aug 01, 2023 |
| 18275145 | TEMPORAL DIFFERENCE SCALING WHEN CONTROLLING AGENTS USING REINFORCEMENT LEARNING | SPRATT, BEAU D | 2143 | §103 | Final Rejection | — | AI Ready | Jul 31, 2023 |
| 18273594 | PREDICTING COMPLETE PROTEIN REPRESENTATIONS FROM MASKED PROTEIN REPRESENTATIONS | LUO, JAMMY NMN | — | §101§103 | Non-Final OA | — | Pending | Jul 21, 2023 |
| 18260182 | PREDICTING EXCHANGE-CORRELATION ENERGIES OF ATOMIC SYSTEMS USING NEURAL NETWORKS | KADING, JOSHUA A | 2121 | §101§102§103§112 | Non-Final OA | — | Pending | Jun 30, 2023 |
| 18267363 | SOLVING MIXED INTEGER PROGRAMS USING NEURAL NETWORKS | ROHD, BENJAMIN MATTHEW | 2147 | §101§103 | Non-Final OA | — | Pending | Jun 14, 2023 |
| 18034989 | PREDICTING PROTEIN STRUCTURES USING PROTEIN GRAPHS | KHAN, ARSHAD HUSSAIN | — | §101§103 | Non-Final OA | — | Pending | May 02, 2023 |
| 18141273 | PRIVACY-SENSITIVE NEURAL NETWORK TRAINING USING DATA AUGMENTATION | HALES, BRIAN J | 2125 | §101§103§112 | Final Rejection | — | AI Ready | Apr 28, 2023 |
| 18034280 | PREDICTING PROTEIN STRUCTURES OVER MULTIPLE ITERATIONS USING RECYCLING | SMITH, JENNIFER JOY | — | §101§102§103§112DP | Non-Final OA | — | Pending | Apr 27, 2023 |
| 18127551 | ALLOCATING COMPUTING RESOURCES BETWEEN MODEL SIZE AND TRAINING DATA DURING TRAINING OF A MACHINE LEARNING MODEL | TAN, DAVID H | 2145 | §102§103 | Final Rejection | 38d overdue | AI Ready | Mar 28, 2023 |
| 18027571 | PREDICTING SYMMETRICAL PROTEIN STRUCTURES USING SYMMETRICAL EXPANSION TRANSFORMATIONS | FONSECA LOPEZ, FRANCINI ALVARENGA | — | §101§103§112 | Non-Final OA | — | Pending | Mar 21, 2023 |
| 18026376 | PREDICTING PROTEIN STRUCTURES BY SHARING INFORMATION BETWEEN MULTIPLE SEQUENCE ALIGNMENTS AND PAIR EMBEDDINGS | BICKHAM, DAWN MARIE | — | §101§102 | Non-Final OA | — | Pending | Mar 15, 2023 |
| 18025689 | TRAINING PROTEIN STRUCTURE PREDICTION NEURAL NETWORKS USING REDUCED MULTIPLE SEQUENCE ALIGNMENTS | HAYES, JONATHAN EDWARD | — | §101§103§112 | Non-Final OA | — | Pending | Mar 10, 2023 |
| 18076978 | TRAINING CONDITIONAL COMPUTATION NEURAL NETWORKS USING REINFORCEMENT LEARNING | GOLAN, MATTHEW BRYCE | 2123 | §103§112 | Final Rejection | — | Pending | Dec 07, 2022 |
| 17960051 | COMPOSITIONAL GENERALIZATION FOR REINFORCEMENT LEARNING | KIM, HARRISON CHAN YOUNG | 2145 | §101§103 | Final Rejection | — | Pending | Oct 04, 2022 |
| 17959232 | NEURAL NETWORKS WITH TRANSFORMED ACTIVATION FUNCTION LAYERS | HADDAD, MAJD MAHER | 2125 | §101§103 | Final Rejection | 31d overdue | AI Ready | Oct 03, 2022 |
| 17798111 | LEARNING MACHINE LEARNING INCENTIVES BY GRADIENT DESCENT FOR AGENT COOPERATION IN A DISTRIBUTED MULTI-AGENT SYSTEM | HAN, KYU HYUNG | 2123 | §103 | Non-Final OA | — | Pending | Aug 08, 2022 |
| 17849269 | DETERMINING A DISTRIBUTION OF ATOM COORDINATES OF A MACROMOLECULE FROM IMAGES USING AUTO-ENCODERS | FONSECA LOPEZ, FRANCINI ALVARENGA | 1685 | §101§102§103 | Final Rejection | — | Pending | Jun 24, 2022 |
| 17763924 | FAST SPARSE NEURAL NETWORKS | MCINTOSH, ANDREW T | 2144 | §103 | Non-Final OA | — | Pending | Mar 25, 2022 |
| 17668050 | SELECTING POINTS IN CONTINUOUS SPACES USING NEURAL NETWORKS | HOOVER, BRENT JOHNSTON | 2100 | §101§103 | Non-Final OA | — | Pending | Feb 09, 2022 |
| 17625361 | TRAINING A NEURAL NETWORK TO CONTROL AN AGENT USING TASK-RELEVANT ADVERSARIAL IMITATION LEARNING | KAPOOR, DEVAN | 2126 | §103 | Final Rejection | 26d | AI Ready | Jan 07, 2022 |
IP Author helps IP teams respond to office actions faster with AI-generated responses, examiner analytics, and prosecution intelligence.
Start Free Trial