3 ms·

I have implemented an AI for credit decisioning in 13 countries on a multi-billion dollar portfolio. Here are my concerns about this elegant yet ineffective pro

I have implemented an AI for credit decisioning in 13 countries on a multi-billion dollar portfolio. Here are my concerns about this elegant yet ineffective prompt: 1. LLMs in general are not build for quantitative analysis. Loan-to-value, income ratios, etc are not supposed to be calculated by such a model. Possible solution would be to calculate this beforehand and provide it to the model or train a submodel using a supervised learning approach to identify good/bad 2. Lending models are governed quarterly yet see relevant cohort changes only after a period of time after credit decision which can be many years. This prompt above does not take this performance of the cohort into consideration 3. Based on the governance companies adjust parameters and models regularly to adjust to changes in the environment. I.e., a new car models comes out or the company is accessing a new customer segment. This process could not be covered well with this prompt since there would be no approvals/ denies for this segment. 4. Since transfer of personal-identififation data needs to be consented, it would likely be necessary to host an LLM like this internally or find a way to ensure there is no data leakage from the provider to other users on the platform. 5. Credit approval limits are not necessarily covered by this proceess. I.e., the credit decisions is unclear but would work with 5-10% more downpayment. Or the customer would be asked to lower the loan value or find someone in the company who can underwrite that loan volume. This person then has usually a bunch of additional questions (liquidity risk, interest risk ,etc) to ensure that the company is well protected and the necessary compliance checks are adhered to. 6. The discussions about this with regulators and auditors will be entertaining. Yet, I think it IS an elegant prompt which might provide some insights.

cs702 4y ago
There's evidence that you can get LLMs to write chain-of-thought explanations that are consistent with the instructions in the given text, including quantitative data, cohort performance, governance imperatives, qualitative considerations, etc. The models can even be given directions to write conditional approvals if necessary. To get a sense of what is and will be possible, take a look at the ReAct paper: https://arxiv.org/abs/2210.03629 https://arxiv.org/abs/2210.03629 and some of the LangChain tutorials that use it: https://langchain.readthedocs.io/en/latest/modules/agents/getting_started.html https://langchain.readthedocs.io/en/latest/modules/agents/ge... https://langchain.readthedocs.io/en/latest/modules/agents/implementations/react.html?highlight=zero-shot%20react https://langchain.readthedocs.io/en/latest/modules/agents/im...
- jsemrau 4y ago
  Thank you for sharing the python code. While I am aware of the paper, I have not seen the code. Let me run a demo with this. Yet overall, this is not only a tech problem, but a compliance / regulatory problem that includes a time differential of sometimes many years. Also, I am not saying its impossible. Mainly because I was the one pushing this type of innovation for a long time and faced headwinds from Credit Operations for many years. Quality prevailed. One comment on the SerpAPI and compliance tools = load_tools(["serpapi", "llm-math"], llm=llm) With this innocuously looking line one integrates Google Search API (through SERPAPI) into the credit approval flow. I.e., you have no control where your customer data might end up with. Second comment: SERPAI sign-up requires email+phone. Why?
  - mark_l_watson 4y ago
    The SerpAPI has backends for a few search engines. And, maybe substitute an ElasticSearch or Nutch search engine? You make a good point.
  - cs702 4y ago
    > Yet overall, this is not only a tech problem, but a compliance / regulatory problem that includes a time differential of sometimes many years. Hmmm... I see what you mean. You're likely right that regulators won't take kindly to this at first. Adoption could take a long while. The difference this time is that executives will be pushing for it!
rattray 4y ago
Can you try it on a few and report back? Would be tremendously curious

cs702 4y ago

There's evidence that you can get LLMs to write chain-of-thought explanations that are consistent with the instructions in the given text, including quantitative data, cohort performance, governance imperatives, qualitative considerations, etc. The models can even be given directions to write conditional approvals if necessary. To get a sense of what is and will be possible, take a look at the ReAct paper: https://arxiv.org/abs/2210.03629 https://arxiv.org/abs/2210.03629 and some of the LangChain tutorials that use it: https://langchain.readthedocs.io/en/latest/modules/agents/getting_started.html https://langchain.readthedocs.io/en/latest/modules/agents/ge... https://langchain.readthedocs.io/en/latest/modules/agents/implementations/react.html?highlight=zero-shot%20react https://langchain.readthedocs.io/en/latest/modules/agents/im...

jsemrau 4y ago
Thank you for sharing the python code. While I am aware of the paper, I have not seen the code. Let me run a demo with this. Yet overall, this is not only a tech problem, but a compliance / regulatory problem that includes a time differential of sometimes many years. Also, I am not saying its impossible. Mainly because I was the one pushing this type of innovation for a long time and faced headwinds from Credit Operations for many years. Quality prevailed. One comment on the SerpAPI and compliance tools = load_tools(["serpapi", "llm-math"], llm=llm) With this innocuously looking line one integrates Google Search API (through SERPAPI) into the credit approval flow. I.e., you have no control where your customer data might end up with. Second comment: SERPAI sign-up requires email+phone. Why?
- mark_l_watson 4y ago
  The SerpAPI has backends for a few search engines. And, maybe substitute an ElasticSearch or Nutch search engine? You make a good point.
- cs702 4y ago
  > Yet overall, this is not only a tech problem, but a compliance / regulatory problem that includes a time differential of sometimes many years. Hmmm... I see what you mean. You're likely right that regulators won't take kindly to this at first. Adoption could take a long while. The difference this time is that executives will be pushing for it!

rattray 4y ago

Can you try it on a few and report back? Would be tremendously curious