Skip to main content

🎉 We released Spotlight 1.6.0 check it out

Version: Next

Detect decision boundaries based on certainty ratios

We use certainty ratios to compute a decision boundary score.

Use Chrome to run Spotlight in Colab. Due to Colab restrictions (e.g. no websocket support), the performance is limited. Run the notebook locally for the full Spotlight experience.

Open In Colab


Imports and play as copy-n-paste functions

# Install dependencies
#@title Install required packages with PIP

!pip install renumics-spotlight datasets
# Play as copy-n-paste functions
import datasets
import numpy as np
import pandas as pd
from renumics import spotlight
import requests

def boundary_score(df, probabilities_name='probabilities'):
def compute_score(probs):
score = [indices[0], indices[1], probs[indices[0]],probs[indices[1]] ]

return score

temp_scores=[compute_score(x) for x in df[probabilities_name]]
df_out['decision_boundary_score']=[x[3]/x[2] for x in temp_scores]
df_out['decision_boundary_alternate_label']=[x[1] for x in temp_scores]

return df_out

Step-by-step example on CIFAR-100

Load CIFAR-100 from Huggingface hub and convert it to Pandas dataframe

dataset = datasets.load_dataset("renumics/cifar100-enriched", split="train")
df = dataset.to_pandas()

Compute decision boundary score and alternate label

f_boundary = boundary_score(df)
df = pd.concat([df, df_boundary], axis=1)
df['decision_boundary_alternate_label_str']=[dataset.features["fine_label"].int2str(x) for x in df['decision_boundary_alternate_label']]

Inspect decision boundary with Spotlight

df_show = df.drop(columns=['embedding', 'probabilities'])
layout_url = ""
response = requests.get(layout_url)
layout = spotlight.layout.nodes.Layout(**json.loads(response.text)), dtype={"image": spotlight.Image, "embedding_reduced": spotlight.Embedding}, layout=layout)