ML+X Community Member Survey (Fall 2023)

JavaScript isn't enabled in your browser, so this file can't be opened. Enable and reload.

Please fill out this brief survey so that we can better understand the backgrounds, practices, and needs of machine learning (ML) practitioners working in Madison. The results from this survey will assist the ML+X community in planning future events.

This survey takes 5 to 10 minutes to complete.

ML-Related Work

In this section we ask you a few questions about your ML-related work

What ML models/methods are you currently learning or using for your work?

How often do you use these programming languages in your ML-related work?

Never

Rarely

Sometimes

Often

Matlab

Python

Julia

Bash

C/C++

JavaScript

Scala

Java

Clear selection

Are there any other languages you sometimes or often use in your ML-related work?

How often do you use the following tools/frameworks in your ML-related work?

Never

Rarely

Sometimes

Often

HuggingFace

NVIDIA tools

PyTorch

Keras

Tensorflow

Scikit-learn (sklearn)

XGBoost

Center for High Throughput Computing (CHTC)

AWS

Google Cloud Platform

Google Colab

Google Copilot

ChatGPT

Microsoft Azure

Clear selection

How often do the following challenges impede your ML-related work?

Never

Rarely

Sometimes

Often

Understanding the data prior to modeling (EDA pipelines)

Diagnosing and improving an ML model

Insufficient computing capabilities (memory, storage space, lack of GPUs, etc.)

Insufficient quantity of data

Insufficient quality of data

Cost of labelling data

Lack of model explanability

Undesirable forms of model bias

Moral concerns (privacy, safety, transparency, bias, potential for harm, ensuring fairness, etc.)

Regulatory concerns (compliance with laws around protected data, auditing, FERPA, HIPAA, IRB, etc.)

ML Library changes / inconsistencies

Difficulty knowing where to begin

Communicating results with stakeholders

Feature selection / feature engineering

Never

Rarely

Sometimes

Often

Understanding the data prior to modeling (EDA pipelines)

Diagnosing and improving an ML model

Insufficient computing capabilities (memory, storage space, lack of GPUs, etc.)

Insufficient quantity of data

Insufficient quality of data

Cost of labelling data

Lack of model explanability

Undesirable forms of model bias

Moral concerns (privacy, safety, transparency, bias, potential for harm, ensuring fairness, etc.)

Regulatory concerns (compliance with laws around protected data, auditing, FERPA, HIPAA, IRB, etc.)

ML Library changes / inconsistencies

Difficulty knowing where to begin

Communicating results with stakeholders

Feature selection / feature engineering

Clear selection

Are there other challenges you sometimes or often face in your ML-related work?

In your ML-related work, how important are the following concerns to you?

Not important

A little important

Important

Very important

The model can be run on limited hardware

The model is not a black-box

The model is fair

The model is well-documented (model use cases, training data sources, model cards, etc.)

The model leads to good impacts

The model's results are interpretable by end-users

The model is accurate

Clear selection

Are their other things important to you in your ML-related work?

Clear form

Never submit passwords through Google Forms.

This form was created inside of UW-Madison Google Workspace. Report Abuse

Forms