1. A computer implemented adaptive likelihood control system (100), the system (100) comprising: a data receiving unit (102) adapted to receive a dataset and corresponding probabilistic contributions associated with each dataset; a posterior construction unit (104) adapted to provide a posterior distribution based on combination of the probabilistic contributions associated with the dataset; a sampling unit (106) adapted to provide sampling data corresponding to estimation of model parameters based on the posterior distribution; a consistency evaluation unit (108) adapted to provide statistical consistency data corresponding to interaction between the probabilistic contributions associated with the dataset; an adaptive control unit (110) adapted to provide modification data corresponding to adjustment of a probabilistic contribution associated with a dataset based on detection of a statistical inconsistency; a processing unit (112) operatively coupled to the data receiving unit (102), the posterior construction unit (104), the sampling unit (106), the consistency evaluation unit (108), and the adaptive control unit (110), characterized in that the processing unit (112) is configured to: construct the posterior distribution based on the probabilistic contributions; execute a sampling procedure using the sampling data for estimation of model parameters; evaluate statistical consistency during execution of the sampling procedure using the statistical consistency data; detect the statistical inconsistency based on a threshold condition; modify the probabilistic contribution associated with the dataset using the modification data; update the posterior distribution based on the modified probabilistic contribution; and continue execution of the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability.

2. The system (100) as claimed in claim 1, wherein the consistency evaluation unit (108) is adapted to provide the statistical consistency data based on a divergence measure, likelihood imbalance, or contribution gradient associated with the probabilistic contributions.

3. The system (100) as claimed in claim 1, wherein the adaptive control unit (110) is adapted to provide the modification data based on a scaling operation, exponentiation operation, attenuation operation, or adaptive weighting operation applied to the probabilistic contribution.

4. The system (100) as claimed in claim 1, wherein the sampling unit (106) is adapted to provide the sampling data using a Markov Chain Monte Carlo technique, a Hamiltonian Monte Carlo technique, a nested sampling technique, or a variational inference technique.

5. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to synchronize the modification of the probabilistic contribution across parallel sampling executions performed on multiple processing environments.

6. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to initialize the probabilistic contributions associated with the dataset with equal influence prior to execution of the sampling procedure.

7. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to dynamically vary the threshold condition for detection of the statistical inconsistency based on progression of the sampling procedure.

8. The system (100) as claimed in claim 1, wherein the posterior construction unit (104) is adapted to update the posterior distribution in real time during execution of the sampling procedure without reinitialization of the sampling procedure.

9. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to reduce exploration of statistically inconsistent parameter regions by controlling influence of the probabilistic contribution associated with the dataset.

10. A computer-implemented method (300) for adaptive likelihood control in multi-dataset probabilistic inference, the method (300) is characterized by steps of: receiving a dataset and corresponding probabilistic contributions associated with each dataset; constructing a posterior distribution by combining the probabilistic contributions associated with the dataset; executing a sampling procedure for estimating model parameters based on the posterior distribution; continuously evaluating, by a processing unit (112) during execution of the sampling procedure, statistical consistency between the probabilistic contributions associated with the dataset using predefined consistency metrics; detecting a statistical inconsistency based on the statistical consistency exceeding a defined threshold condition; adaptively modifying a probabilistic contribution associated with a dataset in response to the detected statistical inconsistency, wherein the modification comprises dynamic adjustment of contribution influence during execution of the sampling procedure; updating the posterior distribution using the modified probabilistic contribution; and continuing the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability. Date: April 08, 2026 Place: Noida Nainsi Rastogi Patent Agent (IN/PA-2372) Agent for the Applicant

Description:BACKGROUND
Field of Invention
[001] Embodiments of the present invention generally relate to a multi-dataset analysis, and computational optimization and particularly to a system for adaptive likelihood control in multi-dataset probabilistic inference.
Description of Related Art
[002] Modern computer-implemented probabilistic inference systems utilize multiple datasets to estimate model parameters; however, such systems face challenges when datasets exhibit statistical inconsistency or conflicting uncertainty characteristics. Fixed likelihood structures assume constant dataset contribution throughout execution, that leads to inefficient processor utilization, excessive computational cycles, unstable convergence, and inaccurate parameter estimation. The absence of automated mechanisms for continuous evaluation of dataset compatibility further aggravates computational inefficiency in multi-dataset environments.
[003] Existing solutions include static Bayesian inference frameworks such as Cobaya, MontePython, and CosmoSIS that combine multiple datasets through predefined likelihood models with fixed weights. Standard sampling techniques such as Markov Chain Monte Carlo, Hamiltonian Monte Carlo, nested sampling, and variational inference operate on static probabilistic structures. Additionally, present practices include manual dataset selection, prior adjustment, and post-analysis diagnostic techniques such as evidence ratios and consistency tests to assess dataset disagreement after computation completes.
[004] Such existing solutions exhibit several shortcomings. Static likelihood integration fails to account for evolving dataset inconsistency, that results in inefficient exploration of parameter space and increased computational cost. Lack of real-time evaluation prevents timely identification of dataset conflict, while dependence on manual intervention reduces reproducibility and increases operational complexity. Post-hoc diagnostic approaches provide only retrospective insights and do not improve computational performance during execution.
[005] There is thus a need for an improved and advanced system and method for adaptive likelihood control in multi-dataset probabilistic inference that can administer the aforementioned limitations in a more efficient manner.
SUMMARY
[006] Embodiments in accordance with the present invention provide a system for adaptive likelihood control in multi-dataset probabilistic inference. The system comprising a data receiving unit adapted to receive a dataset and corresponding probabilistic contributions associated with each dataset. The system further comprising a posterior construction unit adapted to provide a posterior distribution based on combination of the probabilistic contributions associated with the dataset. The system further comprising a sampling unit adapted to provide sampling data corresponding to estimation of model parameters based on the posterior distribution. The system further comprising a consistency evaluation unit adapted to provide statistical consistency data corresponding to interaction between the probabilistic contributions associated with the dataset. The system further comprising an adaptive control unit adapted to provide modification data corresponding to adjustment of a probabilistic contribution associated with a dataset based on detection of a statistical inconsistency. The system further comprising a processing unit operatively coupled to the data receiving unit, the posterior construction unit, the sampling unit, the consistency evaluation unit, and the adaptive control unit. The processing unit is configured to construct the posterior distribution based on the probabilistic contributions, execute a sampling procedure using the sampling data for estimation of model parameters, evaluate statistical consistency during execution of the sampling procedure using the statistical consistency data, detect the statistical inconsistency based on a threshold condition, modify the probabilistic contribution associated with the dataset using the modification data, update the posterior distribution based on the modified probabilistic contribution, and continue execution of the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability.
[007] Embodiments in accordance with the present invention further provide a method for adaptive likelihood control in multi-dataset probabilistic inference. The method comprising steps of: receiving a dataset and corresponding probabilistic contributions associated with each dataset; constructing a posterior distribution by combining the probabilistic contributions associated with the dataset; executing a sampling procedure for estimating model parameters based on the posterior distribution; continuously evaluating, by a processing unit during execution of the sampling procedure, statistical consistency between the probabilistic contributions associated with the dataset using predefined consistency metrics; detecting a statistical inconsistency based on the statistical consistency exceeding a defined threshold condition; adaptively modifying a probabilistic contribution associated with a dataset in response to the detected statistical inconsistency, wherein the modification comprises dynamic adjustment of contribution influence during execution of the sampling procedure; updating the posterior distribution using the modified probabilistic contribution; and continuing the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability.
[008] Embodiments of the present invention may provide a number of advantages depending on their particular configuration. First, embodiments of the present application may provide a system for adaptive likelihood control in multi-dataset probabilistic inference
[009] Next, embodiments of the present application may provide a system for adaptive likelihood control in multi-dataset probabilistic inference that improves computational efficiency by reducing redundant likelihood evaluations and minimizing unnecessary processor cycles during probabilistic inference.
[0010] Next, embodiments of the present application may provide a system for adaptive likelihood control in multi-dataset probabilistic inference that enhances convergence stability by mitigating the impact of statistically inconsistent datasets on parameter estimation.
[0011] Next, embodiments of the present application may provide a system for adaptive likelihood control in multi-dataset probabilistic inference that reduces dependence on manual intervention by enabling automated handling of dataset inconsistencies within the inference process.
[0012] Next, embodiments of the present application may provide a system for adaptive likelihood control in multi-dataset probabilistic inference that optimizes resource utilization by limiting exploration of incompatible parameter regions and improving overall processing performance.
[0013] These and other advantages will be apparent from the present application of the embodiments described herein.
[0014] The preceding is a simplified summary to provide an understanding of some embodiments of the present invention. This summary is neither an extensive nor exhaustive overview of the present invention and its various embodiments. The summary presents selected concepts of the embodiments of the present invention in a simplified form as an introduction to the more detailed description presented below. As will be appreciated, other embodiments of the present invention are possible utilizing, alone or in combination, one or more of the features set forth above or described in detail below.
BRIEF DESCRIPTION OF THE DRAWINGS
[0015] The above and still further features and advantages of embodiments of the present invention will become apparent upon consideration of the following detailed description of embodiments thereof, especially when taken in conjunction with the accompanying drawings, and wherein:
[0016] FIG. 1 illustrates block diagram of system for adaptive likelihood control in multi-dataset probabilistic inference, according to an embodiment of the present invention;
[0017] FIG. 2 illustrates components of a processing unit of the system for adaptive likelihood control in multi-dataset probabilistic inference, according to an embodiment of the present invention; and
[0018] FIG. 3 depicts a flowchart of a method for adaptive likelihood control in multi-dataset probabilistic inference, according to an embodiment of the present invention.
[0019] The headings used herein are for organizational purposes only and are not meant to be used to limit the scope of the description or the claims. As used throughout this application, the word "may" is used in a permissive sense (i.e., meaning having the potential to), rather than the mandatory sense (i.e., meaning must). Similarly, the words “include”, “including”, and “includes” mean including but not limited to. To facilitate understanding, like reference numerals have been used, where possible, to designate like elements common to the figures. Optional portions of the figures may be illustrated using dashed or dotted lines, unless the context of usage indicates otherwise.

DETAILED DESCRIPTION
[0020] The following description includes the preferred best mode of one embodiment of the present invention. It will be clear from this description of the invention that the invention is not limited to these illustrated embodiments but that the invention also includes a variety of modifications and embodiments thereto. Therefore, the present description should be seen as illustrative and not limiting. While the invention is susceptible to various modifications and alternative constructions, it should be understood, that there is no intention to limit the invention to the specific form disclosed, but, on the contrary, the invention is to cover all modifications, alternative constructions, and equivalents falling within the scope of the invention as defined in the claims.
[0021] In any embodiment described herein, the open-ended terms "comprising", "comprises”, and the like (which are synonymous with "including", "having” and "characterized by") may be replaced by the respective partially closed phrases "consisting essentially of", “consists essentially of", and the like or the respective closed phrases "consisting of", "consists of”, the like.
[0022] As used herein, the singular forms “a”, “an”, and “the” designate both the singular and the plural, unless expressly stated to designate the singular only.
[0023] FIG. 1 illustrates a block diagram of a system 100 for adaptive likelihood control in multi-dataset probabilistic inference, according to an embodiment of the present application. The system 100 may be robust, scalable, and adaptive to dynamically regulate probabilistic contributions, manage inference execution, and provide improved computational efficiency and convergence stability during parameter estimation. The system 100 may provide a closed-loop computational architecture for adaptive regulation of probabilistic contributions during probabilistic inference. Thus, improving computational efficiency, convergence stability, and scalability across diverse computing environments.
[0024] In an embodiment of the present invention, the system 100 may be adapted to detect instability in parameter estimation based on fluctuations in probabilistic contributions or divergence in sampling trajectories. The system 100 may be further adapted to mitigate instability by reducing influence of conflicting datasets and stabilizing posterior updates. The system 100 may maintain consistent convergence behaviour during probabilistic inference.
[0025] In an embodiment of the present invention, unlike existing systems that operate using static probabilistic structures, the system 100 may be adapted to dynamically modify probabilistic contributions during execution of the sampling procedure based on real-time statistical evaluation. The integration of continuous evaluation and adaptive modification within the same computational workflow may provide a technical improvement in processing efficiency and convergence stability, that may not be achievable through static likelihood integration or post-hoc diagnostic approaches.
[0026] According to the embodiments of the present invention, the system 100 may incorporate non-limiting hardware components to enhance the processing speed and efficiency such as the system 100 may comprise a data receiving unit 102, a posterior construction unit 104, a sampling unit 106, a consistency evaluation unit 108, an adaptive control unit 110, and a processing unit 112. In an embodiment of the present invention, the hardware components of the system 100 may be integrated with computer-executable instructions for overcoming the challenges and the limitations of the existing systems.
[0027] In an embodiment of the present invention, the data receiving unit 102 may be adapted to receive a dataset and corresponding probabilistic contributions associated with each dataset. The datasets may comprise structured data, unstructured data, observational data, experimental data, simulated data, and so forth. The probabilistic contributions may include likelihood functions, probability density functions, equivalent probabilistic representations, and so forth. The data receiving unit 102 may be, but not limited to, a data interface, memory interface, communication interface, and so forth adapted to obtain datasets from local storage, distributed systems, remote computing environments, and so forth. Embodiments of the present application are intended to include or otherwise cover any type of datasets and probabilistic contributions, including known, related art, and/or later developed technologies.
[0028] The data receiving unit 102 may be, but not limited to, a data interface, memory interface, communication interface, and so forth. Embodiments of the present invention are intended to include or otherwise cover any type of the data receiving unit 102, including known, related art, and/or later developed technologies.
[0029] In an embodiment of the present invention, the posterior construction unit 104 may be adapted to provide a posterior distribution based on combination of the probabilistic contributions associated with the dataset. The posterior distribution may represent a composite probabilistic model used for parameter estimation. The posterior construction unit 104 may combine the probabilistic contributions using predefined probabilistic formulations. Embodiments of the present application are intended to include or otherwise cover any type of posterior construction techniques, including known, related art, and/or later developed technologies.
[0030] The posterior construction unit 104 may be, but not limited to, a probabilistic computation engine, statistical modelling unit, and so forth. Embodiments of the present invention are intended to include or otherwise cover any type of the posterior construction unit 104, including known, related art, and/or later developed technologies.
[0031] In an embodiment of the present invention, the sampling unit 106 may be adapted to provide sampling data corresponding to estimation of model parameters based on the posterior distribution. The sampling data may comprise parameter samples, likelihood evaluations, transition states, and so forth associated with exploration of a parameter space. The sampling unit 106 may be, but not limited to, a processor-executed sampling engine configured to implement sampling techniques such as Markov Chain Monte Carlo, Hamiltonian Monte Carlo, nested sampling, variational inference, and so forth. Embodiments of the present application are intended to include or otherwise cover any type of sampling techniques, including known, related art, and/or later developed technologies.
[0032] The sampling unit 106 may be, but not limited to, a sampling engine, stochastic simulation unit, and so forth. Embodiments of the present invention are intended to include or otherwise cover any type of the sampling unit 106, including known, related art, and/or later developed technologies.
[0033] In an embodiment of the present invention, the consistency evaluation unit 108 may be adapted to provide statistical consistency data corresponding to interaction between the probabilistic contributions associated with the dataset. The statistical consistency data may comprise divergence measures, likelihood imbalance, contribution gradients, equivalent statistical indicators, and so forth. The consistency evaluation unit 108 may be configured to determine compatibility or inconsistency among datasets during execution of a sampling procedure. Embodiments of the present application are intended to include or otherwise cover any type of statistical consistency evaluation techniques, including known, related art, and/or later developed technologies.
[0034] In an embodiment of the present invention, the consistency evaluation unit 108 may be adapted to compute multiple statistical consistency metrics simultaneously and may combine the metrics using a weighted aggregation mechanism. The weighted aggregation mechanism may assign dynamic weights to individual metrics based on reliability, sensitivity, or stage of sampling procedure. The consistency evaluation unit 108 may further be adapted to select or prioritize specific metrics during execution of the sampling procedure to improve accuracy of statistical inconsistency detection.
[0035] In an embodiment of the present invention, the consistency evaluation unit 108 may be adapted to model interaction between datasets by analysing relative influence of probabilistic contributions across parameter space. The consistency evaluation unit 108 may compute pairwise interaction measures, aggregated interaction measures, or hierarchical interaction representations to determine statistical relationships among datasets. The interaction modelling may enable identification of conflicting dataset behaviour, reinforcing dataset contributions exhibiting agreement and attenuating dataset contributions exhibiting inconsistency.
[0036] The consistency evaluation unit 108 may be, but not limited to, a statistical analysis unit, divergence computation unit, and so forth. Embodiments of the present invention are intended to include or otherwise cover any type of the consistency evaluation unit 108, including known, related art, and/or later developed technologies.
[0037] In an embodiment of the present invention, the adaptive control unit 110 may be adapted to provide modification data corresponding to adjustment of a probabilistic contribution associated with a dataset based on detection of a statistical inconsistency. The modification data may comprise scaling parameters, exponentiation factors, attenuation coefficients, adaptive weights, and so forth applied to the probabilistic contribution. The adaptive control unit 110 may regulate influence of datasets based on detected statistical inconsistency. Embodiments of the present application are intended to include or otherwise cover any type of adaptive modification operations, including known, related art, and/or later developed technologies.
[0038] In an embodiment of the present invention, the adaptive control unit 110 may be adapted to select a modification strategy based on characteristics of detected statistical inconsistency. The adaptive control unit 110 may determine whether scaling, exponentiation, attenuation, or adaptive weighting is to be applied based on magnitude, persistence, or type of inconsistency. The adaptive control unit 110 may further be adapted to switch between modification strategies during execution of the sampling procedure to maintain stability and computational efficiency.
[0039] The adaptive control unit 110 may be, but not limited to, a control logic unit, adaptive weighting unit, and so forth. Embodiments of the present invention are intended to include or otherwise cover any type of the adaptive control unit 110, including known, related art, and/or later developed technologies.
[0040] In an embodiment of the present invention, the processing unit 112 may be operatively coupled to the data receiving unit 102, the posterior construction unit 104, the sampling unit 106, the consistency evaluation unit 108, and the adaptive control unit 110. The processing unit 112 may be, but not limited to, processors, microprocessors, controllers, computing devices configured to execute instructions stored in a memory. The processing unit 112 may be configured to construct the posterior distribution based on the probabilistic contributions and execute the sampling procedure using the sampling data for estimation of model parameters. The processing unit 112 may further evaluate statistical consistency during execution of the sampling procedure using the statistical consistency data and detect statistical inconsistency based on the threshold condition.
[0041] In an embodiment of the present invention, the threshold condition may be dynamically adjusted by the processing unit 112 based on progression of the sampling procedure. The processing unit 112 may be adapted to modify the threshold condition using convergence indicators, sampling variance, or stability metrics associated with probabilistic contributions. The dynamic threshold adjustment may enable early detection of strong inconsistencies and refined detection during later stages of convergence.
[0042] In an embodiment of the present invention, the processing unit 112 may be adapted to determine convergence based on a dual convergence condition comprising convergence of model parameter estimates and convergence of adaptive modification parameters associated with probabilistic contributions. The processing unit 112 may evaluate stability of parameter distributions and stability of modification data simultaneously. The sampling procedure may be terminated only when both convergence conditions are satisfied within predefined tolerance limits, thereby ensuring stable and reliable inference.
[0043] In an embodiment of the present invention, the processing unit 112 may be adapted to perform evaluation of statistical consistency and adaptive modification of probabilistic contributions within each iteration of the sampling procedure. The evaluation and modification operations may be tightly coupled with iterative sampling steps such that probabilistic contributions may be dynamically updated during traversal of parameter space. The processing unit 112 may ensure that adaptive modification is executed without termination or reinitialization of the sampling procedure, thereby enabling in-loop correction of dataset influence.
[0044] In an embodiment of the present invention, upon detection of statistical inconsistency, the processing unit 112 may utilize the modification data to adjust the probabilistic contribution associated with the dataset. The processing unit 112 may update the posterior distribution based on the modified probabilistic contribution and continue execution of the sampling procedure using the updated posterior distribution without reinitialization of the sampling procedure.
[0045] In an embodiment of the present invention, the processing unit 112 may be configured to initialize probabilistic contributions associated with the dataset with equal influence prior to execution of the sampling procedure. The processing unit 112 may dynamically vary the threshold condition for detection of statistical inconsistency based on progression of the sampling procedure. The processing unit 112 may synchronize modification of probabilistic contributions across parallel sampling executions performed on multiple processing environments to ensure consistency in distributed inference systems. The posterior construction unit 104 may update the posterior distribution in real time during execution of the sampling procedure without reinitialization. The processing unit 112 may reduce exploration of statistically inconsistent parameter regions by controlling influence of probabilistic contributions.
[0046] In an embodiment of the present invention, the processing unit 112 may be adapted to coordinate synchronization of modification data across distributed processing environments using a consistency preservation mechanism. The processing unit 112 may ensure that updated probabilistic contributions are propagated across parallel sampling executions while maintaining coherence of posterior distribution. The synchronization mechanism may be adapted to resolve conflicts between parallel updates and maintain consistency of adaptive modifications.
[0047] In an embodiment of the present invention, the processing unit 112 may be adapted to reduce exploration of statistically inconsistent parameter regions by attenuating influence of conflicting probabilistic contributions. The reduction in exploration may lead to decreased number of sampling iterations, reduced likelihood evaluations, and optimized utilization of processing resources. The system 100 may achieve improved computational efficiency through controlled adaptation of dataset influence.
[0048] In an embodiment of the present invention, the processing unit 112 may be adapted to optimize processor utilization and memory access by limiting redundant computations associated with inconsistent datasets. The adaptive modification of probabilistic contributions may reduce unnecessary likelihood evaluations and memory operations. The system 100 may improve processing throughput and reduce computational overhead in multi-dataset inference environments.
[0049] The processing unit 112 may be, but not limited to, a processor, microprocessor, controller, computing device, and so forth. Embodiments of the present invention are intended to include or otherwise cover any type of the processing unit 112, including known, related art, and/or later developed technologies. The processing unit 112 may further be explained in detail in conjunction with FIG. 2.
[0050] In an embodiment of the present invention, the system 100 may be adapted to implement an explicit closed-loop control mechanism. The consistency evaluation unit 108 may generate statistical consistency data that may be continuously provided to the adaptive control unit 110, and the adaptive control unit 110 may generate modification data that may be applied to probabilistic contributions through the processing unit 112. The processing unit 112 may be further adapted to update the posterior distribution based on the modification data and may re-initiate evaluation through the consistency evaluation unit 108, thereby forming a continuous feedback loop. The closed-loop control mechanism may ensure iterative refinement of probabilistic contributions based on real-time statistical behaviour of datasets during execution of the sampling procedure.
[0051] FIG. 2 illustrates components of the processing unit 112 of the system 100, according to an embodiment of the present application. The processing unit 112 may comprise a posterior management module 200, a sampling control module 202, a consistency monitoring module 204, an adaptive modification module 206, and a parallel synchronization module 208. The modules may collectively enable construction, evaluation, modification, and execution of probabilistic inference in a closed-loop manner.
[0052] In an embodiment of the present invention, the posterior management module 200 may be configured to construct the posterior distribution based on probabilistic contributions associated with the dataset. The posterior management module 200 may be further configured to receive probabilistic contribution data from the posterior construction unit 104. The posterior management module 200 may be configured to update the posterior distribution upon modification of probabilistic contributions. Further, the posterior management module 200 may be configured to transmit updated posterior data to the sampling control module 202.
[0053] In an embodiment of the present invention, the sampling control module 202 may be configured to receive the updated posterior data from the posterior management module 200. The sampling control module 202 may be further configured to control execution of the sampling procedure for estimation of model parameters. The sampling control module 202 may be configured to manage iterative parameter estimation cycles based on the posterior distribution. In an exemplary scenario, the sampling control module 202 may be configured to continue execution of the sampling procedure after update of the posterior distribution without reinitialization. Further, the sampling control module 202 may be configured to transmit sampling progression data to the consistency monitoring module 204.
[0054] In an embodiment of the present invention, the consistency monitoring module 204 may be configured to receive statistical consistency data from the consistency evaluation unit 108. The consistency monitoring module 204 may be configured to evaluate interaction between probabilistic contributions associated with the dataset. The consistency monitoring module 204 may be further configured to detect statistical inconsistency based on the predefined or dynamically varying threshold condition. In an exemplary scenario, if the consistency monitoring module 204 determines that statistical inconsistency exceeds the threshold condition, then the consistency monitoring module 204 may be configured to generate an inconsistency signal. Further, the consistency monitoring module 204 may be configured to transmit the inconsistency signal to the adaptive modification module 206.
[0055] In an embodiment of the present invention, the adaptive modification module 206 may be configured to receive the inconsistency signal from the consistency monitoring module 204 and modification data from the adaptive control unit 110. The adaptive modification module 206 may be configured to modify probabilistic contributions associated with the datasets. The modification may include, but not limited to, scaling, attenuation, exponentiation, adaptive weighting, and so forth of the probabilistic contributions. In an exemplary scenario, the adaptive modification module 206 may be configured to generate modified probabilistic contribution data and transmit the modified probabilistic contribution data to the posterior management module 200 for updating the posterior distribution.
[0056] In an embodiment of the present invention, the parallel synchronization module 208 may be configured to coordinate execution of probabilistic inference across multiple processing environments. The parallel synchronization module 208 may be configured to receive modification data and sampling progression data from the adaptive modification module 206 and the sampling control module 202. The parallel synchronization module 208 may be further configured to synchronize modification of probabilistic contributions across parallel sampling executions. Further, the parallel synchronization module 208 may be configured to ensure consistency of posterior updates across distributed processing unit(s) 112 by coordinating data exchange between multiple processing nodes.
[0057] FIG. 3 depicts a flowchart of a method 300 for adaptive likelihood control in multi-dataset probabilistic inference, according to an embodiment of the present invention.
[0058] At step 302, the system 100 may receive the dataset and corresponding probabilistic contributions associated with each dataset.
[0059] At step 304, the system 100 may construct the posterior distribution by combining the probabilistic contributions associated with the dataset.
[0060] At step 306, the system 100 may execute the sampling procedure for estimating the model parameters based on the posterior distribution.
[0061] At step 308, the system 100 may continuously evaluate, the sampling procedure, the statistical consistency between the probabilistic contributions associated with the dataset using the predefined consistency metrics.
[0062] At step 310, the system 100 may detect the statistical inconsistency based on the statistical consistency exceeding the defined threshold condition.
[0063] At step 312, the system 100 may adaptively modify the probabilistic contribution associated with the dataset in response to the detected statistical inconsistency. The modification comprises dynamic adjustment of contribution influence during execution of the sampling procedure.
[0064] At step 314, the system 100 may update the posterior distribution using the modified probabilistic contribution.
[0065] At step 316, the system 100 may continue the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability.
[0066] While the invention has been described in connection with what is presently considered to be the most practical and various embodiments, it is to be understood that the invention is not to be limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims.
[0067] This written description uses examples to disclose the invention, including the best mode, and also to enable any person skilled in the art to practice the invention, including making and using any devices or systems and performing any incorporated methods. The patentable scope of the invention is defined in the claims, and may include other examples that occur to those skilled in the art. Such other examples are intended to be within the scope of the claims if they have structural elements that do not differ from the literal language of the claims, or if they include equivalent structural elements within substantial differences from the literal languages of the claims. , Claims:CLAIMS
I/We Claim:
1. A computer implemented adaptive likelihood control system (100), the system (100) comprising:
a data receiving unit (102) adapted to receive a dataset and corresponding probabilistic contributions associated with each dataset;
a posterior construction unit (104) adapted to provide a posterior distribution based on combination of the probabilistic contributions associated with the dataset;
a sampling unit (106) adapted to provide sampling data corresponding to estimation of model parameters based on the posterior distribution;
a consistency evaluation unit (108) adapted to provide statistical consistency data corresponding to interaction between the probabilistic contributions associated with the dataset;
an adaptive control unit (110) adapted to provide modification data corresponding to adjustment of a probabilistic contribution associated with a dataset based on detection of a statistical inconsistency;
a processing unit (112) operatively coupled to the data receiving unit (102), the posterior construction unit (104), the sampling unit (106), the consistency evaluation unit (108), and the adaptive control unit (110), characterized in that the processing unit (112) is configured to:
construct the posterior distribution based on the probabilistic contributions;
execute a sampling procedure using the sampling data for estimation of model parameters;
evaluate statistical consistency during execution of the sampling procedure using the statistical consistency data;
detect the statistical inconsistency based on a threshold condition;
modify the probabilistic contribution associated with the dataset using the modification data;
update the posterior distribution based on the modified probabilistic contribution; and
continue execution of the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability.
2. The system (100) as claimed in claim 1, wherein the consistency evaluation unit (108) is adapted to provide the statistical consistency data based on a divergence measure, likelihood imbalance, or contribution gradient associated with the probabilistic contributions.
3. The system (100) as claimed in claim 1, wherein the adaptive control unit (110) is adapted to provide the modification data based on a scaling operation, exponentiation operation, attenuation operation, or adaptive weighting operation applied to the probabilistic contribution.
4. The system (100) as claimed in claim 1, wherein the sampling unit (106) is adapted to provide the sampling data using a Markov Chain Monte Carlo technique, a Hamiltonian Monte Carlo technique, a nested sampling technique, or a variational inference technique.
5. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to synchronize the modification of the probabilistic contribution across parallel sampling executions performed on multiple processing environments.
6. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to initialize the probabilistic contributions associated with the dataset with equal influence prior to execution of the sampling procedure.
7. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to dynamically vary the threshold condition for detection of the statistical inconsistency based on progression of the sampling procedure.
8. The system (100) as claimed in claim 1, wherein the posterior construction unit (104) is adapted to update the posterior distribution in real time during execution of the sampling procedure without reinitialization of the sampling procedure.
9. The system (100) as claimed in claim 1, wherein the processing unit (112) is configured to reduce exploration of statistically inconsistent parameter regions by controlling influence of the probabilistic contribution associated with the dataset.
10. A computer-implemented method (300) for adaptive likelihood control in multi-dataset probabilistic inference, the method (300) is characterized by steps of:
receiving a dataset and corresponding probabilistic contributions associated with each dataset;
constructing a posterior distribution by combining the probabilistic contributions associated with the dataset;
executing a sampling procedure for estimating model parameters based on the posterior distribution;
continuously evaluating, by a processing unit (112) during execution of the sampling procedure, statistical consistency between the probabilistic contributions associated with the dataset using predefined consistency metrics;
detecting a statistical inconsistency based on the statistical consistency exceeding a defined threshold condition;
adaptively modifying a probabilistic contribution associated with a dataset in response to the detected statistical inconsistency, wherein the modification comprises dynamic adjustment of contribution influence during execution of the sampling procedure;
updating the posterior distribution using the modified probabilistic contribution; and
continuing the sampling procedure using the updated posterior distribution to improve computational efficiency and convergence stability.
Date: April 08, 2026
Place: Noida

Nainsi Rastogi
Patent Agent (IN/PA-2372)
Agent for the Applicant