Delete detailedResults.json with huggingface_hub

detailedResults.json

@@ -1 +0,0 @@
-[{"": "1", "Id": "0", "Fullsentence": "accordinglythere is a need to connect the organizational level ai governanceneeds into ai implementation processes across the solution lifecycledue to the increasing presence of ml models in software systems data scientist and ml engineers have become prominent rolesin software development teams", "Component": "connect the", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "2", "Id": "0", "Fullsentence": "accordinglythere is a need to connect the organizational level ai governanceneeds into ai implementation processes across the solution lifecycledue to the increasing presence of ml models in software systems data scientist and ml engineers have become prominent rolesin software development teams", "Component": "organizational level ai governanceneeds", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "5", "Id": "0", "Fullsentence": "accordinglythere is a need to connect the organizational level ai governanceneeds into ai implementation processes across the solution lifecycledue to the increasing presence of ml models in software systems data scientist and ml engineers have become prominent rolesin software development teams", "Component": "implementation processes across the solution lifecycle", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "7", "Id": "0", "Fullsentence": "accordinglythere is a need to connect the organizational level ai governanceneeds into ai implementation processes across the solution lifecycledue to the increasing presence of ml models in software systems data scientist and ml engineers have become prominent rolesin software development teams", "Component": "increasing presence of ml models in software systems data scientist and ml engineers have become prominent rolesin software development teams", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "8", "Id": "1", "Fullsentence": "this is not a straightforwardtask as the process of creating ml models is often experimentaland takes place in uncharted territory and is consequently moredifficult to predict than more mundane well established it systemdevelopment processes  2022", "Component": "this is not a straightforwardtask as", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "10", "Id": "1", "Fullsentence": "this is not a straightforwardtask as the process of creating ml models is often experimentaland takes place in uncharted territory and is consequently moredifficult to predict than more mundane well established it systemdevelopment processes  2022", "Component": "creating ml models is often experimentaland takes place in uncharted territory", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "12", "Id": "1", "Fullsentence": "this is not a straightforwardtask as the process of creating ml models is often experimentaland takes place in uncharted territory and is consequently moredifficult to predict than more mundane well established it systemdevelopment processes  2022", "Component": "moredifficult to predict than more mundane well established it systemdevelopment processes 2022", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "14", "Id": "2", "Fullsentence": "furthermore due to the characteristic of ml models being inscrutable black boxes issues such asensuring explainability  2 and establishing audit trails for the mlmodels  912 have to be accounted for", "Component": "the characteristic of ml models being inscrutable black boxes", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "15", "Id": "2", "Fullsentence": "furthermore due to the characteristic of ml models being inscrutable black boxes issues such asensuring explainability  2 and establishing audit trails for the mlmodels  912 have to be accounted for", "Component": "issues such asensuring explainability 2", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "16", "Id": "2", "Fullsentence": "furthermore due to the characteristic of ml models being inscrutable black boxes issues such asensuring explainability  2 and establishing audit trails for the mlmodels  912 have to be accounted for", "Component": "establishing audit trails for the mlmodels 912 have to be accounted for", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "17", "Id": "3", "Fullsentence": "our study advances the understanding ofai governance in the context of ml model development and thuscontributes to the emerging body of is literature on ai systemdesign and development  35", "Component": "study advances the understanding ofai governance in the context of ml model development", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "20", "Id": "4", "Fullsentence": "this process follows thecontinuous integrationcontinuous development cicd paradigmwhich implies that as code is pushed into a repository it is automatically and continuously integrated into the surrounding systemand deployed into production after a series of automated tests andchecks  2952", "Component": "process follows thecontinuous integrationcontinuous development cicd paradigmwhich implies", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "21", "Id": "4", "Fullsentence": "this process follows thecontinuous integrationcontinuous development cicd paradigmwhich implies that as code is pushed into a repository it is automatically and continuously integrated into the surrounding systemand deployed into production after a series of automated tests andchecks  2952", "Component": "pushed into a repository", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "22", "Id": "4", "Fullsentence": "this process follows thecontinuous integrationcontinuous development cicd paradigmwhich implies that as code is pushed into a repository it is automatically and continuously integrated into the surrounding systemand deployed into production after a series of automated tests andchecks  2952", "Component": "it is automatically and continuously integrated into the surrounding systemand deployed into production after", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "25", "Id": "4", "Fullsentence": "this process follows thecontinuous integrationcontinuous development cicd paradigmwhich implies that as code is pushed into a repository it is automatically and continuously integrated into the surrounding systemand deployed into production after a series of automated tests andchecks  2952", "Component": "tests andchecks 2952", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "26", "Id": "5", "Fullsentence": "still this paradigm can be seen asthe current industry standard and is encouraged by cloud vendorssuch as google cloud  22 and used by major online services suchas facebook 14the growing applicability of ml models for solving various taskshas also led to the emergence of new developer roles in softwaredevelopment teams in particular that of a data scientist  49", "Component": "paradigm can be seen asthe current industry standard", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "28", "Id": "5", "Fullsentence": "still this paradigm can be seen asthe current industry standard and is encouraged by cloud vendorssuch as google cloud  22 and used by major online services suchas facebook 14the growing applicability of ml models for solving various taskshas also led to the emergence of new developer roles in softwaredevelopment teams in particular that of a data scientist  49", "Component": "applicability of ml models for solving various tasks", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "30", "Id": "5", "Fullsentence": "still this paradigm can be seen asthe current industry standard and is encouraged by cloud vendorssuch as google cloud  22 and used by major online services suchas facebook 14the growing applicability of ml models for solving various taskshas also led to the emergence of new developer roles in softwaredevelopment teams in particular that of a data scientist  49", "Component": "emergence of new developer roles in softwaredevelopment teams in particular that of a data scientist 49", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "32", "Id": "6", "Fullsentence": "thus work integrating the work ofdata scientists into sdlc models is neededstudies have acknowledged that ml systems require appropriatesoftware engineering workflows and verification and validationtesting  43", "Component": "work of  c", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "34", "Id": "6", "Fullsentence": "thus work integrating the work ofdata scientists into sdlc models is neededstudies have acknowledged that ml systems require appropriatesoftware engineering workflows and verification and validationtesting  43", "Component": "needed  udi", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "37", "Id": "6", "Fullsentence": "thus work integrating the work ofdata scientists into sdlc models is neededstudies have acknowledged that ml systems require appropriatesoftware engineering workflows and verification and validationtesting  43", "Component": "acknowledged that ml systems require appropriatesoftware engineering workflows", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "41", "Id": "6", "Fullsentence": "thus work integrating the work ofdata scientists into sdlc models is neededstudies have acknowledged that ml systems require appropriatesoftware engineering workflows and verification and validationtesting  43", "Component": "validationtesting 43", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "43", "Id": "7", "Fullsentence": "due to theway ml systems are built and operate the human operators seethe input and output values but do not necessarily have an explicitunderstanding of how the algorithm processes the data  2", "Component": "theway ml systems are built and operate", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "44", "Id": "7", "Fullsentence": "due to theway ml systems are built and operate the human operators seethe input and output values but do not necessarily have an explicitunderstanding of how the algorithm processes the data  2", "Component": "the human operators seethe input and output values but do not necessarily have an explicitunderstanding of how the algorithm processes the data 2", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "46", "Id": "8", "Fullsentence": "thus system engineers are forced to work with confidence intervals and particularlythe more complex models such as dnns may never reach 100prediction accuracy 1third ml model development is datadriven", "Component": "system engineers", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "47", "Id": "8", "Fullsentence": "thus system engineers are forced to work with confidence intervals and particularlythe more complex models such as dnns may never reach 100prediction accuracy 1third ml model development is datadriven", "Component": "forced to", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "48", "Id": "8", "Fullsentence": "thus system engineers are forced to work with confidence intervals and particularlythe more complex models such as dnns may never reach 100prediction accuracy 1third ml model development is datadriven", "Component": "work with confidence intervals", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "49", "Id": "8", "Fullsentence": "thus system engineers are forced to work with confidence intervals and particularlythe more complex models such as dnns may never reach 100prediction accuracy 1third ml model development is datadriven", "Component": "more complex models such as dnns may never reach 100prediction accuracy 1third ml model development is datadriven", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "51", "Id": "9", "Fullsentence": "thus data scientists have emerged  49", "Component": "data scientists have emerged 49", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "55", "Id": "10", "Fullsentence": "this data can thenbe used to train new even more holistic modelsfourth ml models cannot be tweaked once trained and thusnew models need to be created to replace them", "Component": "more holistic", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "56", "Id": "10", "Fullsentence": "this data can thenbe used to train new even more holistic modelsfourth ml models cannot be tweaked once trained and thusnew models need to be created to replace them", "Component": "ml models cannot be tweaked once trained", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "59", "Id": "11", "Fullsentence": "governance of data sources is therefore paramount toensure holistic ai governance 41modelrelated ai governance is nontrivial due to complex mlmodels having poor explainability  2", "Component": "governance of data sources", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "60", "Id": "11", "Fullsentence": "governance of data sources is therefore paramount toensure holistic ai governance 41modelrelated ai governance is nontrivial due to complex mlmodels having poor explainability  2", "Component": "paramount toensure holistic ai governance 41modelrelated ai governance is nontrivial", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "62", "Id": "11", "Fullsentence": "governance of data sources is therefore paramount toensure holistic ai governance 41modelrelated ai governance is nontrivial due to complex mlmodels having poor explainability  2", "Component": "complex mlmodels having poor explainability 2", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "63", "Id": "12", "Fullsentence": "during the process of connecting our 2nd order themes into the two aggregate dimensionswe noticed that many of the 2nd order themes were antecedents tothe devops process", "Component": "order themes", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "65", "Id": "12", "Fullsentence": "during the process of connecting our 2nd order themes into the two aggregate dimensionswe noticed that many of the 2nd order themes were antecedents tothe devops process", "Component": "many of the 2nd order themes were ant  ecede", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "69", "Id": "13", "Fullsentence": "hence the very first thing a project should lookinto is the suitability of ml for solving the business problem", "Component": "project should", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "71", "Id": "13", "Fullsentence": "hence the very first thing a project should lookinto is the suitability of ml for solving the business problem", "Component": "is the suitability of ml for solving the business problem", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "73", "Id": "14", "Fullsentence": "thefollowing quotes stress this pointwhen applying machine learning or ai the most central anddifficult thing is to find the business case", "Component": "most central anddifficult thing", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "74", "Id": "14", "Fullsentence": "thefollowing quotes stress this pointwhen applying machine learning or ai the most central anddifficult thing is to find the business case", "Component": "to find the business case  scope what data is related to the case", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "77", "Id": "15", "Fullsentence": "an example quote from one of the informantsdiscussing the matter is presented belownowadays when starting a project with a customer we alwaysscope what data is related to the case and if a strong data perspectiveis there we involve a data scientist to the project for further inquiries", "Component": "involve a data scientist to the project for further inquiries", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "78", "Id": "16", "Fullsentence": "model interpretation as complexity grows and there are so many parameters at some point noone is able to understand what is going on", "Component": "model interpretation as complexity grows", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "79", "Id": "16", "Fullsentence": "model interpretation as complexity grows and there are so many parameters at some point noone is able to understand what is going on", "Component": "there are so many parameters at some point", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "80", "Id": "16", "Fullsentence": "model interpretation as complexity grows and there are so many parameters at some point noone is able to understand what is going on", "Component": "able to understand what is going on", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "81", "Id": "17", "Fullsentence": "this sentiment was popular among the informantsby contrast a few informants brought up that relying on cloudvendors may not be possible in all scenarios due to data securityissues", "Component": "relying on cloudvendors may not be possible in all scenarios", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "83", "Id": "17", "Fullsentence": "this sentiment was popular among the informantsby contrast a few informants brought up that relying on cloudvendors may not be possible in all scenarios due to data securityissues", "Component": "data securityissues", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "84", "Id": "18", "Fullsentence": "all these three aspects have intrinsicai governance aspects to considerthe participants gave examples of several systems that they hadbeen part of creating where ml models are injected and connectedto and consequently made part of the larger system", "Component": "three aspects have intrinsicai governance aspects", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "86", "Id": "18", "Fullsentence": "all these three aspects have intrinsicai governance aspects to considerthe participants gave examples of several systems that they hadbeen part of creating where ml models are injected and connectedto and consequently made part of the larger system", "Component": "gave examples of several systems that they hadbeen part of creating where ml models are injected and connectedto", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "88", "Id": "18", "Fullsentence": "all these three aspects have intrinsicai governance aspects to considerthe participants gave examples of several systems that they hadbeen part of creating where ml models are injected and connectedto and consequently made part of the larger system", "Component": "made part of the larger system", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "94", "Id": "19", "Fullsentence": "during the interviews libraries such as tensorflow kerasand pytorch surfaced many times and the cloud environments towhich for example tensorflow is tied to were mentioned as wellas data storage and computation oftentimes additionally takes placein these cloud platforms their role with in the system design isenormous", "Component": "cloud platforms their role with in the system design isenormous", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "96", "Id": "20", "Fullsentence": "informant15 explainedin particular in machine learning to produce results that can bereplicated we need data versioning there are tools for this  forexample dvc3", "Component": "there are tools for this forexample dvc3", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "97", "Id": "21", "Fullsentence": "furthermore data validation wasdeemed important due to privacy policies and related regulatoryrequirements", "Component": "data validation wasdeemed important", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "99", "Id": "21", "Fullsentence": "furthermore data validation wasdeemed important due to privacy policies and related regulatoryrequirements", "Component": "privacy policies and related regulatoryrequirements", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "100", "Id": "22", "Fullsentence": "when multiple models are trained a wayto govern the process is linking the model version number to thedataset it was trained withmodel development was seen as a highly iterative process andduring this process many things related to the algorithms and training approach change", "Component": "multiple models are trained", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "101", "Id": "22", "Fullsentence": "when multiple models are trained a wayto govern the process is linking the model version number to thedataset it was trained withmodel development was seen as a highly iterative process andduring this process many things related to the algorithms and training approach change", "Component": "model version number  dataset", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "103", "Id": "22", "Fullsentence": "when multiple models are trained a wayto govern the process is linking the model version number to thedataset it was trained withmodel development was seen as a highly iterative process andduring this process many things related to the algorithms and training approach change", "Component": "it was trained withmodel development was seen as a highly iterative process  during this process many things related to the algorithms and training approach change", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "105", "Id": "23", "Fullsentence": "according to the informants most projects involving ml require a greatdeal of experimenting and testingwhen training a model we realize that the data we have does notfill the expectations we set for it or our original idea does not workand we need to go back", "Component": "projects involving m", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "106", "Id": "23", "Fullsentence": "according to the informants most projects involving ml require a greatdeal of experimenting and testingwhen training a model we realize that the data we have does notfill the expectations we set for it or our original idea does not workand we need to go back", "Component": "require a greatdeal of experimenting and testingwhen training a model we realize that the data we have does notfill the expectations we set for it or our original idea does not work  and we need to go back", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "108", "Id": "24", "Fullsentence": "reinforcement learningby contrast does not typically use much external data but theenvironment in which the reinforcement learning agent is beingtrained needs to be trackedthe informants pointed out that parameter tweaking is a majorpart of getting the model to work as intended but that there existsvarious tools to assist in and automate this process", "Component": "reinforcement learning", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "109", "Id": "24", "Fullsentence": "reinforcement learningby contrast does not typically use much external data but theenvironment in which the reinforcement learning agent is beingtrained needs to be trackedthe informants pointed out that parameter tweaking is a majorpart of getting the model to work as intended but that there existsvarious tools to assist in and automate this process", "Component": "not typically use much external data", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "110", "Id": "24", "Fullsentence": "reinforcement learningby contrast does not typically use much external data but theenvironment in which the reinforcement learning agent is beingtrained needs to be trackedthe informants pointed out that parameter tweaking is a majorpart of getting the model to work as intended but that there existsvarious tools to assist in and automate this process", "Component": "parameter tweaking", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "111", "Id": "24", "Fullsentence": "reinforcement learningby contrast does not typically use much external data but theenvironment in which the reinforcement learning agent is beingtrained needs to be trackedthe informants pointed out that parameter tweaking is a majorpart of getting the model to work as intended but that there existsvarious tools to assist in and automate this process", "Component": "a majorpart of getting the model to work as intended", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "112", "Id": "24", "Fullsentence": "reinforcement learningby contrast does not typically use much external data but theenvironment in which the reinforcement learning agent is beingtrained needs to be trackedthe informants pointed out that parameter tweaking is a majorpart of getting the model to work as intended but that there existsvarious tools to assist in and automate this process", "Component": "there existsvarious tools to assist", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "113", "Id": "24", "Fullsentence": "reinforcement learningby contrast does not typically use much external data but theenvironment in which the reinforcement learning agent is beingtrained needs to be trackedthe informants pointed out that parameter tweaking is a majorpart of getting the model to work as intended but that there existsvarious tools to assist in and automate this process", "Component": "in and automate this process", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "114", "Id": "25", "Fullsentence": "once the model startssprouting out something quite wild its good there is someone whoquickly stops the process to avoid further damage", "Component": "startssprouting out something quite wild  quick", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "116", "Id": "25", "Fullsentence": "once the model startssprouting out something quite wild its good there is someone whoquickly stops the process to avoid further damage", "Component": "the process", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "118", "Id": "25", "Fullsentence": "once the model startssprouting out something quite wild its good there is someone whoquickly stops the process to avoid further damage", "Component": "avoid further damage", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "121", "Id": "26", "Fullsentence": "theaccuracy is thus 97 period", "Component": "97 period", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "122", "Id": "27", "Fullsentence": "here biasesrefer to unfair predictions made by the model and anomalies tounusual occurrences and events in the ai system output or theoverall systemthe importance of automated monitoring from the governanceperspective is critical as it acts as a final layer of security in systemswhere ml model predictions need to be accurate", "Component": "biases  unusual occurrences", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "125", "Id": "27", "Fullsentence": "here biasesrefer to unfair predictions made by the model and anomalies tounusual occurrences and events in the ai system output or theoverall systemthe importance of automated monitoring from the governanceperspective is critical as it acts as a final layer of security in systemswhere ml model predictions need to be accurate", "Component": "ai system output or theoverall system", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "126", "Id": "27", "Fullsentence": "here biasesrefer to unfair predictions made by the model and anomalies tounusual occurrences and events in the ai system output or theoverall systemthe importance of automated monitoring from the governanceperspective is critical as it acts as a final layer of security in systemswhere ml model predictions need to be accurate", "Component": "importance of automated monitoring from the governanceperspective is critical as it acts as a final layer of security in systemswhere ml model", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "127", "Id": "27", "Fullsentence": "here biasesrefer to unfair predictions made by the model and anomalies tounusual occurrences and events in the ai system output or theoverall systemthe importance of automated monitoring from the governanceperspective is critical as it acts as a final layer of security in systemswhere ml model predictions need to be accurate", "Component": "predictions need to be accurate", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "129", "Id": "28", "Fullsentence": "hencethe findings can revisualized as a sdlc model", "Component": "findings can revisualized as a sdlc model", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "130", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "contributions  ai governance 35", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "132", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "early steps for the conceptualunification of ai governance", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "133", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "sdlc models", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "134", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "research gap on implementing ai governance in sdlc  but", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "136", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "systematic mapping of", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "137", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "technical ai", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "138", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "governance dimensions to common sdlc models", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "139", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "sequential waterfalltypeapproach", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "140", "Id": "29", "Fullsentence": "with these contributions we advance research onai governance  35 by providing early steps for the conceptualunification of ai governance and sdlc modelswith reference to the research gap on implementing ai governance in sdlc models identified in the introduction our workcontributes a systematic mapping of technical ai governance dimensions to common sdlc models a sequential waterfalltypeapproach and a devopsbased cicd approach", "Component": "a de  vo", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "143", "Id": "30", "Fullsentence": "to complementprevious exploratory work by researchers  34 our empirical studyis based on interviews of highprofile ai experts and thus firmlygrounded in software development practice", "Component": "complementprevious exploratory work by researchers 34 our empirical studyis based on interviews of highprofile ai experts", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "145", "Id": "30", "Fullsentence": "to complementprevious exploratory work by researchers  34 our empirical studyis based on interviews of highprofile ai experts and thus firmlygrounded in software development practice", "Component": "firmlygrounded in software development practice", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "146", "Id": "31", "Fullsentence": "the foundation of technical ai governance ultimately enables beneficial ai development and use onthe organizational and societal layers 1553 contributions to practiceprevious work has stressed that when software consulting companies deliver ml systems the final product should be the mlopspipeline  48", "Component": "the foundation of technical ai governance ultimately enables", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "147", "Id": "31", "Fullsentence": "the foundation of technical ai governance ultimately enables beneficial ai development and use onthe organizational and societal layers 1553 contributions to practiceprevious work has stressed that when software consulting companies deliver ml systems the final product should be the mlopspipeline  48", "Component": "beneficial  i", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "150", "Id": "31", "Fullsentence": "the foundation of technical ai governance ultimately enables beneficial ai development and use onthe organizational and societal layers 1553 contributions to practiceprevious work has stressed that when software consulting companies deliver ml systems the final product should be the mlopspipeline  48", "Component": "practiceprevious work", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "151", "Id": "31", "Fullsentence": "the foundation of technical ai governance ultimately enables beneficial ai development and use onthe organizational and societal layers 1553 contributions to practiceprevious work has stressed that when software consulting companies deliver ml systems the final product should be the mlopspipeline  48", "Component": "stressed that", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "152", "Id": "31", "Fullsentence": "the foundation of technical ai governance ultimately enables beneficial ai development and use onthe organizational and societal layers 1553 contributions to practiceprevious work has stressed that when software consulting companies deliver ml systems the final product should be the mlopspipeline  48", "Component": "when software consulting companies deliver ml systems", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "153", "Id": "31", "Fullsentence": "the foundation of technical ai governance ultimately enables beneficial ai development and use onthe organizational and societal layers 1553 contributions to practiceprevious work has stressed that when software consulting companies deliver ml systems the final product should be the mlopspipeline  48", "Component": "the final product should be the mlopspipeline 48", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "155", "Id": "32", "Fullsentence": "due to the heterogeneity and continuously changing and evolving nature of aisystems it is important to bring clarity to ai governance with conceptualizations and visualizations that can be applied broadly toseveral projectswhile previous work on the technical layer of ai governancehas neatly described it in the three themes of data model andsystem  41 our findings offer a more rugged viewpoint on thetopic", "Component": "the heterogeneity and continuously changing and evolving nature of aisystems", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "156", "Id": "32", "Fullsentence": "due to the heterogeneity and continuously changing and evolving nature of aisystems it is important to bring clarity to ai governance with conceptualizations and visualizations that can be applied broadly toseveral projectswhile previous work on the technical layer of ai governancehas neatly described it in the three themes of data model andsystem  41 our findings offer a more rugged viewpoint on thetopic", "Component": "important to bring clarity to ai governance with conceptualizations and visualizations that can be applied broadly toseveral projectswhile previous work on the technical layer of ai governancehas neatly described it in", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "157", "Id": "32", "Fullsentence": "due to the heterogeneity and continuously changing and evolving nature of aisystems it is important to bring clarity to ai governance with conceptualizations and visualizations that can be applied broadly toseveral projectswhile previous work on the technical layer of ai governancehas neatly described it in the three themes of data model andsystem  41 our findings offer a more rugged viewpoint on thetopic", "Component": "our findings offer a more rugged viewpoint on thetopic", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "158", "Id": "33", "Fullsentence": "altogether we feel that we had a sufficient saturation of interviews atthis stage and see more detailed field experiments case studies andsystemspecific analyses as a promising direction to continue thisline of researchfor future research the results of this work could also be furtherfinetuned by conducting a delphistudy  28 where the findingspresented in this work are sent back to the interviewed experts oranother group of experts and critically evaluated based on receivedfeedback", "Component": "that we had a sufficient saturation of interviews atthis stage", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "159", "Id": "33", "Fullsentence": "altogether we feel that we had a sufficient saturation of interviews atthis stage and see more detailed field experiments case studies andsystemspecific analyses as a promising direction to continue thisline of researchfor future research the results of this work could also be furtherfinetuned by conducting a delphistudy  28 where the findingspresented in this work are sent back to the interviewed experts oranother group of experts and critically evaluated based on receivedfeedback", "Component": "more detailed field experiments case studies andsystemspecific analyses as a promising direction to", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "160", "Id": "33", "Fullsentence": "altogether we feel that we had a sufficient saturation of interviews atthis stage and see more detailed field experiments case studies andsystemspecific analyses as a promising direction to continue thisline of researchfor future research the results of this work could also be furtherfinetuned by conducting a delphistudy  28 where the findingspresented in this work are sent back to the interviewed experts oranother group of experts and critically evaluated based on receivedfeedback", "Component": "results of this work could also be furtherfinetuned by conducting a delphistudy 28", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "161", "Id": "33", "Fullsentence": "altogether we feel that we had a sufficient saturation of interviews atthis stage and see more detailed field experiments case studies andsystemspecific analyses as a promising direction to continue thisline of researchfor future research the results of this work could also be furtherfinetuned by conducting a delphistudy  28 where the findingspresented in this work are sent back to the interviewed experts oranother group of experts and critically evaluated based on receivedfeedback", "Component": "findingspresented in this work", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "162", "Id": "33", "Fullsentence": "altogether we feel that we had a sufficient saturation of interviews atthis stage and see more detailed field experiments case studies andsystemspecific analyses as a promising direction to continue thisline of researchfor future research the results of this work could also be furtherfinetuned by conducting a delphistudy  28 where the findingspresented in this work are sent back to the interviewed experts oranother group of experts and critically evaluated based on receivedfeedback", "Component": "sent back to the interviewed experts oranother group of experts", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "163", "Id": "33", "Fullsentence": "altogether we feel that we had a sufficient saturation of interviews atthis stage and see more detailed field experiments case studies andsystemspecific analyses as a promising direction to continue thisline of researchfor future research the results of this work could also be furtherfinetuned by conducting a delphistudy  28 where the findingspresented in this work are sent back to the interviewed experts oranother group of experts and critically evaluated based on receivedfeedback", "Component": "critically evaluated based on receivedfeedback", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "164", "Id": "34", "Fullsentence": "hence we see it fruitful to take even more detailed andgranular perspectives towards organizational ai governance andthen ultimately connecting them back to sdlc conceptualizationsand consequently concrete development work6 conclusionsin this work we conducted indepth expert interviews to elucidatethe governance needs of technical ml development in view ofthe sdlc", "Component": "we see it fruitful to take even more detailed andgranular perspectives towards organizational ai governance  the", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "166", "Id": "34", "Fullsentence": "hence we see it fruitful to take even more detailed andgranular perspectives towards organizational ai governance andthen ultimately connecting them back to sdlc conceptualizationsand consequently concrete development work6 conclusionsin this work we conducted indepth expert interviews to elucidatethe governance needs of technical ml development in view ofthe sdlc", "Component": "ultimately connecting them back to sdlc conceptualizations", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "168", "Id": "34", "Fullsentence": "hence we see it fruitful to take even more detailed andgranular perspectives towards organizational ai governance andthen ultimately connecting them back to sdlc conceptualizationsand consequently concrete development work6 conclusionsin this work we conducted indepth expert interviews to elucidatethe governance needs of technical ml development in view ofthe sdlc", "Component": "concrete development work6 conclusionsin this work we conducted indepth expert interviews to elucidatethe governance needs of technical ml development in view ofthe sdlc", "causeOrEffect": "effect", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "169", "Id": "35", "Fullsentence": "for example the ongoingresearch on xai tools  2 as a potential solution to governance issues arising from the black box nature of dnns and other complexai systems", "Component": "ongoing  res", "causeOrEffect": "cause", "Labellevel1": "Performance", "Labellevel2": "Investors"}, {"": "173", "Id": "35", "Fullsentence": "for example the ongoingresearch on xai tools  2 as a potential solution to governance issues arising from the black box nature of dnns and other complexai systems", "Component": "a potential solution to governance issues arising from the black", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "174", "Id": "35", "Fullsentence": "for example the ongoingresearch on xai tools  2 as a potential solution to governance issues arising from the black box nature of dnns and other complexai systems", "Component": "box nature of dnns and other complexai systems", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "175", "Id": "36", "Fullsentence": "currently it seems that the sheer complexity of these models prevents any attempts to gain insights intohow they operate and as deep neural networks convolutional networks and other complex ml techniques are advancing the modelexplainability may only worse over time", "Component": "sheer complexity of these models prevents", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "176", "Id": "36", "Fullsentence": "currently it seems that the sheer complexity of these models prevents any attempts to gain insights intohow they operate and as deep neural networks convolutional networks and other complex ml techniques are advancing the modelexplainability may only worse over time", "Component": "attempts to gain insights intohow they operate", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "177", "Id": "36", "Fullsentence": "currently it seems that the sheer complexity of these models prevents any attempts to gain insights intohow they operate and as deep neural networks convolutional networks and other complex ml techniques are advancing the modelexplainability may only worse over time", "Component": "neural networks con  volution", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "180", "Id": "36", "Fullsentence": "currently it seems that the sheer complexity of these models prevents any attempts to gain insights intohow they operate and as deep neural networks convolutional networks and other complex ml techniques are advancing the modelexplainability may only worse over time", "Component": "other complex ml", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "182", "Id": "36", "Fullsentence": "currently it seems that the sheer complexity of these models prevents any attempts to gain insights intohow they operate and as deep neural networks convolutional networks and other complex ml techniques are advancing the modelexplainability may only worse over time", "Component": "advancing the modelexplainability may only worse over time", "causeOrEffect": "effect", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}, {"": "183", "Id": "37", "Fullsentence": "trends and trajectories in the software industry implicationsfor the future of work", "Component": "trends and trajectories in the software industry implicationsfo  r the future of work", "causeOrEffect": "cause", "Labellevel1": "Non-Performance", "Labellevel2": "Non-performance"}]