Where are operators of "Degree of Plausible" and "Get GDP and Population"?

fangkuoyu

Dr. Ingo mentioned two operators "Degree of Plausible" and "Get GDP and Population" in his talk "No Way Back: Machine Learning in the Age of Generative AI". However, I cannot find these operators in RapidMiner 10.3. Does anyone know where they are? 

Thanks

Frank

RolandJones

Hi Frank,

The operators used are all part of the Generative Models extension, specifically the Send Prompt (OpenAI). Yes, it is likely with any public model that the answer will potentially be different or vary with time. If you would want consistent results, then you would likely need to have your own private model available without these limitations. Altair RapidMiner offers users the opportunity to finetune models so you would then be able to get the consistent results that you would want.

Best,

Roland

RolandJones

Hi @fangkuoyu,

I believe that these aren't standalone operators, rather operators from the Generative Models extension being used for different activities. Given their use, I would suspect they are using the Open AI model type.

Hope this helps,

Best,

Roland

fangkuoyu

rjones13, 

Thanks for your comments. I have attached the screen dumps from the talk below. I cannot find those operators in Generative Models Extension. Just wonder if they are under other extensions or other operator names. It doesn't look like a direct response from a chat model under OpenAI.

Regards
Frank

RolandJones

Hi Frank,

I've spoken to Ingo, and indeed these operators are using the OpenAI method for getting this information. 

I've shown a small example below of how it would work, but you will need to provide your own OpenAI API key. You can find more information here: https://docs.rapidminer.com/latest/studio/generative-ai/#connecting-to-openai

Best,

Roland

<?xml version="1.0" encoding="UTF-8"?><process version="10.3.001">

  <context>

    <input/>

    <output/>

    <macros/>

  </context>

  <operator activated="true" class="process" compatibility="10.3.001" expanded="true" name="Process" origin="GENERATED_TUTORIAL">

    <parameter key="logverbosity" value="init"/>

    <parameter key="random_seed" value="2001"/>

    <parameter key="send_mail" value="never"/>

    <parameter key="notification_email" value=""/>

    <parameter key="process_duration_for_mail" value="30"/>

    <parameter key="encoding" value="SYSTEM"/>

    <process expanded="true">

      <operator activated="true" class="utility:create_exampleset" compatibility="10.3.001" expanded="true" height="68" name="Create ExampleSet" origin="GENERATED_TUTORIAL" width="90" x="45" y="34">

        <parameter key="generator_type" value="comma separated text"/>

        <parameter key="number_of_examples" value="5"/>

        <parameter key="use_stepsize" value="false"/>

        <list key="function_descriptions"/>

        <parameter key="add_id_attribute" value="false"/>

        <list key="numeric_series_configuration"/>

        <list key="date_series_configuration"/>

        <list key="date_series_configuration (interval)"/>

        <parameter key="date_format" value="yyyy-MM-dd HH:mm:ss"/>

        <parameter key="time_zone" value="SYSTEM"/>

        <parameter key="input_csv_text" value="Country&#10;Germany&#10;Croatia&#10;Belgium"/>

        <parameter key="column_separator" value=","/>

        <parameter key="parse_all_as_nominal" value="false"/>

        <parameter key="decimal_point_character" value="."/>

        <parameter key="trim_attribute_names" value="true"/>

      </operator>

      <operator activated="true" class="retrieve" compatibility="10.3.001" expanded="true" height="68" name="Retrieve" origin="GENERATED_TUTORIAL" width="90" x="45" y="187">

        <description align="center" color="transparent" colored="false" width="126">IMPORTANT: you will need to provide your API key as a Dictionary Connection to make this work. Provide this connection as the second input.</description>

      </operator>

      <operator activated="true" class="generative_models:send_prompt_openai" compatibility="2.0.000" expanded="true" height="82" name="Get GDP" origin="GENERATED_TUTORIAL" width="90" x="179" y="34">

        <parameter key="model" value="gpt-3.5-turbo"/>

        <parameter key="name" value="GDP"/>

        <parameter key="prompt" value="What is the GDP of [[Country]]? Just the answer, no prose."/>

        <parameter key="max_target_tokens" value="256"/>

        <parameter key="temperature" value="1.0"/>

        <parameter key="top_p" value="1.0"/>

        <parameter key="frequency_penalty" value="0.0"/>

        <parameter key="presence_penalty" value="0.0"/>

        <parameter key="conda_environment" value="rm_genai"/>

        <description align="center" color="transparent" colored="false" width="126">The result of this operator will be a data set enriched with the results of the prompts sent to OpenAI.</description>

      </operator>

      <connect from_op="Create ExampleSet" from_port="output" to_op="Get GDP" to_port="data"/>

      <connect from_op="Retrieve" from_port="output" to_op="Get GDP" to_port="connection"/>

      <connect from_op="Get GDP" from_port="data" to_port="result 1"/>

      <portSpacing port="source_input 1" spacing="0"/>

      <portSpacing port="sink_result 1" spacing="0"/>

      <portSpacing port="sink_result 2" spacing="0"/>

    </process>

  </operator>

</process>

fangkuoyu

rjones13:

Thanks for the process. Is it in the official release under some extensions or an experimental operator/process? Bytheway, I think that the model might return different values on different trials. If so, the process/operator might get inconsistent results over time.

RolandJones

Hi Frank,

The operators used are all part of the Generative Models extension, specifically the Send Prompt (OpenAI). Yes, it is likely with any public model that the answer will potentially be different or vary with time. If you would want consistent results, then you would likely need to have your own private model available without these limitations. Altair RapidMiner offers users the opportunity to finetune models so you would then be able to get the consistent results that you would want.

Best,

Roland

fangkuoyu

rjones13:

The "Get GDP" operator is implemented by the "Send Prompt" operator. Now I go it. Thanks for your comments!

Regards
Frank