Let’s look at one more interesting thing. I mentioned before that one of the purposes of this example was to examine and visualize our eigenfaces, as they are called: our super columns. I will not disappoint. Let’s write some code that will show us our super columns as they would look to us humans: def […]
Understanding pre-trained models Pre-trained models are like learning from the experience of others. These models have been trained on extensive datasets, learning patterns, and features that make them adept at their tasks. Think of it as if a model has been reading thousands of books on a subject, absorbing all that information. When we use […]
Example – TL for image classification We could take a pre-trained model such as ResNet or the Vision Transformer (shown in Figure 12.10), initially trained on a large-scale image dataset such as ImageNet. This model has already learned to detect various features from images, from simple shapes to complex objects. We can take advantage of […]
Measuring bias To successfully combat bias, we must first measure its existence and understand its impact on our ML models. Several statistical methods and techniques have been developed for this purpose, each offering a different perspective on bias and fairness. Here are a few essential methods: These methods offer diverse perspectives on measuring bias and […]
Mitigation during model in-processing So, we’ve done our best to clean up our data, but what about when we’re training our model? Can we do something there too? Absolutely! The model in-processing stage allows us to bake fairness right into the training process. It’s a bit like adding spices to a dish as it cooks, […]
Uncovering bias in GPT-2 LLMs such as GPT-2 have revolutionized natural language processing (NLP) by producing high-quality, human-like text. While these models can be valuable tools, they are not without their challenges. One significant concern is the potential for bias in their outputs, which can result from biases in the training data. To understand this […]
Mitigating drift The world of ML is ever-evolving, making it crucial for us to remain adaptable. We’ve seen how the concept of drift is integral to understanding changes in our data or model over time. But what can we do when faced with these shifting sands? Are we merely left to witness the disintegration of […]
Mastering data governance For countless organizations, data stands as a priceless treasure. Yet, it’s data governance that serves as the compass guiding one to extract the true worth of data. Envision data governance as a holistic amalgamation of principles, methodologies, and tools designed to oversee the entire life cycle of your data, ensuring it’s in […]
Data integration – from collection to delivery Data, in its raw form, originates from a myriad of sources, demanding integration to unlock its full potential. The integration might unfurl in batch sequences or be streamed in real time, aiming for prompt insight derivation. But integration isn’t a standalone task—it demands orchestrated actions to curate, ingest, […]
We can then see the visual interpretation (Figure 14.1) of which tokens impact the final decision the most: # Sample tweet to explaintweet = “I love using the new feature!So helpful.”# Generate the explanationexp = explainer.explain_instance(tweet, predictor, num_features=5, top_labels=3)exp.show_in_notebook() In this case, num_features determines how many features the explainer should use to describe the prediction. […]
