New definition of interpretability makes model design more actionable.
problem Current definitions of interpretability are not actionable and inform users poorly.
method Proposes a new definition of interpretability that is general, simple, and actionable.
result New definition reveals necessary properties for designing interpretable models.
New method improves neural network interpretability against adversarial attacks.
problem Adversarial attacks can hide from neural network interpretability methods.
method Develops an interpretability-aware defensive scheme promoting robust interpretation.
result Achieves both robust classification and robust interpretation.
VALC provides concept-level interpretations of FLMs, overcoming word-level limitations.
problem Lack of higher-level structure interpretation in FLMs' attention weights.
method Formal definition of conceptual interpretation, variational Bayesian framework (VALC).
result VALC finds optimal language concepts for FLM predictions, providing concept-level interpretations.
New framework improves interpretability of trainable prompts.
problem Improving task-specific LLM performance with soft prompts remains a black-box method.
method Developed a theoretical framework for evaluating interpretability of trainable prompts, inspired new objective functions.
result Found a fundamental trade-off between interpretability and task performance in trainable prompts.
Interpreting machine learning models helps understand adversarial attacks and defenses.
problem Understanding model vulnerability to adversarial attacks.
method Model interpretation techniques to explore adversarial attacks and defenses.
result Interpretation methods can be applied to adversarial attacks and defenses.
There is a need of ensuring machine learning models that are interpretable. Higher interpretability of the model means easier comprehension and explanation of future predictions for end-users. Further, interpretable machine learning models allow healthcare experts to make reasonable and data-driven decisions to provide…
Model-agnostic interpretation methods can mislead if not used carefully.
problem Misinterpretation of machine learning models due to improper use of techniques.
method General pitfalls of model-agnostic interpretation methods.
result Many pitfalls exist when using global interpretation techniques for machine learning models.
Study on trade-offs between accuracy and interpretability in machine learning.
problem Lack of formal study on statistical cost of interpretability.
method Modeling interpretability as a constraint in empirical risk minimization for binary classification.
result Explains conditions under which accuracy trade-off occurs with interpretability constraints.
Supervised machine learning models boast remarkable predictive capabilities. But can you trust your model? Will it work in deployment? What else can it tell you about the world? We want models to be not only good, but interpretable. And yet the task of interpretation appears underspecified. Papers provide diverse and s…
New framework learns interpretable rule ensembles without sacrificing accuracy.
problem Trade-off between accuracy and interpretability in rule ensembles.
method Introduces local interpretability and a regularizer to promote it, using coordinate descent with local search.
result Learns rule ensembles with fewer rules to explain individual predictions, maintaining comparable accuracy.
Study finds machine learning interpretations are often unstable and unreliable.
problem Reliability of machine learning interpretations in high-stakes domains.
method Stability study on global interpretations using tabular data.
result Popular interpretation methods are frequently unstable, less stable than predictions, and not associated with prediction accuracy.
Meta-learning approach to learn interpretable models from human feedback.
problem Tackling the challenge of making machine learning models interpretable.
method A meta-learning approach where a model of non-trivial proxies of human interpretability is learned from human feedback, then incorporated into the ML training process to optimize for interpretability.
result The approach leads to formulas that are either significantly more or equally accurate while being more interpretable.
A salient approach to interpretable machine learning is to restrict modeling to simple models. In the Bayesian framework, this can be pursued by restricting the model structure and prior to favor interpretable models. Fundamentally, however, interpretability is about users' preferences, not the data generation mechanis…
We create interpretable word embeddings through sparse coding.
problem Difficult to interpret word embeddings in natural language processing.
method Transform pretrained dense word embeddings into sparse embeddings through sparse coding.
result Sparse embeddings are more interpretable and achieve good performance.
Proposes a model to interpret complex ML algorithms.
problem Complex ML models are hard to interpret.
method Uses model-based regression trees and interpretable main-effects models.
result Surrogate model provides interpretable results with good predictive performance.
We propose to use boosted regression trees as a way to compute human-interpretable solutions to reinforcement learning problems. Boosting combines several regression trees to improve their accuracy without significantly reducing their inherent interpretability. Prior work has focused independently on reinforcement lear…
We often desire our models to be interpretable as well as accurate. Prior work on optimizing models for interpretability has relied on easy-to-quantify proxies for interpretability, such as sparsity or the number of operations required. In this work, we optimize for interpretability by directly including humans in the …
As machine learning systems become ubiquitous, there has been a surge of interest in interpretable machine learning: systems that provide explanation for their outputs. These explanations are often used to qualitatively assess other criteria such as safety or non-discrimination. However, despite the interest in interpr…
New method learns to encode predictions within interpretations, improving evaluation.
problem Need for interpretable machine learning, but existing methods are slow or lack fidelity.
method Amortized explanation methods that learn a global selector model optimizing fidelity of interpretations.
result Predictions can be encoded within interpretations, detected by EVAL-X.
GMT improves interpretability of XGNNs by approximating SubMT.
problem Limited understanding of existing interpretable subgraph learning methods.
method Formulated subgraph multilinear extension (SubMT) and designed GMT architecture.
result GMT outperforms state-of-the-art in both interpretability and generalizability.
Most research on the interpretability of machine learning systems focuses on the development of a more rigorous notion of interpretability. I suggest that a better understanding of the deficiencies of the intuitive notion of interpretability is needed as well. I show that visualization enables but also impedes intuitiv…
A framework learns interpretable models with minimal accuracy loss.
problem Tackles interpretability in deep learning models.
method Jointly learns a predictive model and its interpretation model with minimal loss of accuracy.
result Generates interpretable models with minimal loss of accuracy.
InterpretML is an open-source Python package which exposes machine learning interpretability algorithms to practitioners and researchers. InterpretML exposes two types of interpretability - glassbox models, which are machine learning models designed for interpretability (ex: linear models, rule lists, generalized addit…
Machine-generated interpretations do not improve users' guessing accuracy in image classifiers.
problem Determining the usefulness of machine-generated explanations for deep neural networks.
method Human evaluation of crowd workers guessing incorrectly predicted labels with and without visual interpretations.
result Showing machine-generated visual interpretations decreased average guessing accuracy by about 10%.
There is an interpretation of open string field theory in algebraic topology. An interpretation of closed string field theory can be deduced from this open string theory to obtain as well the interpretation of open and closed string field theory combined.
Shallow trees in ensemble models make models more interpretable and sometimes better.
problem Lack of transparency in high-performing tree ensemble models.
method Developed an interpretation algorithm to convert tree ensembles into functional ANOVA representations. Proposed strategies to enhance interpretability.
result Shallow trees in ensemble models can lead to better generalization performance and improved interpretability.
Interpretable neural networks improve economic research by balancing accuracy and transparency.
problem Lack of interpretability in neural networks hinders their use in economic research.
method Proposes interpretable neural network models that balance prediction accuracy and interpretability.
result Achieved 94.5% accuracy in predicting employment status using high-dimensional data.
QUACKIE creates a new benchmark for NLP interpretability.
problem Evaluating NLP interpretability methods is challenging due to biased ground truths.
method Formulated a custom classification task from question-answering datasets, generating unbiased ground truths.
result Demonstrated the effectiveness of current interpretability methods on the new benchmark.
Researchers review challenges in interpreting additive models, especially neural additive models.
problem Challenges in interpreting additive models, particularly neural additive models.
method Review of generalized additive models and discussion of nonidentifiability.
result Challenges in claiming interpretability or suitability for safety-critical applications of additive models.
The paper connects decision tree interpretability and robustness through separation.
problem Empirical observation of a connection between robustness and interpretability in decision trees.
method Investigation of the connection through decision trees and l∞-perturbation robustness, proving bounds on tree size. result First algorithm with guarantees on robustness, interpretability, and accuracy for decision trees.
LI-ITR combines flexible ML with interpretable approximations for personalized treatment rules.
problem Combining flexibility and interpretability in personalized treatment rules.
method Uses variational autoencoders and a mixture of interpretable experts.
result Accurately recovers true local coefficients and optimal treatment strategies.
The paper connects neural networks to Mahalanobis distance for interpretability.
problem Lack of interpretability in neural networks.
method Establishes a connection between neural network linear layers and Mahalanobis distance.
result Provides a foundation for more interpretable neural network models.
Geometrically interprets frequency in electric circuits.
problem Conventional frequency definition limitations.
method Introduces a multivector definition of frequency.
result Conventional frequency is a special case of the new framework.
Understanding black-box machine learning models is crucial for their widespread adoption. Learning globally interpretable models is one approach, but achieving high performance with them is challenging. An alternative approach is to explain individual predictions using locally interpretable models. For locally interpre…
RCAV quantifies model sensitivity to semantic concepts, improving interpretability methods.
problem Lack of semantic interpretability in image classification models.
method RCAV calculates concept gradients and ascent steps to assess model sensitivity to semantic concepts.
result RCAV yields more accurate and robust interpretations of model behavior.
DeepCoDA provides personalized interpretability for complex health data.
problem Interpreting complex health data, especially compositional data, is challenging.
method DeepCoDA framework for high-dimensional compositional data, personalized interpretability through patient-specific weights.
result DeepCoDA maintains state-of-the-art performance and provides coherent, personalized interpretations.
Interpretable deep learning is a fundamental building block towards safer AI, especially when the deployment possibilities of deep learning-based computer-aided medical diagnostic systems are so eminent. However, without a computational formulation of black-box interpretation, general interpretability research rely hea…
Deep learning interpretation is essential to explain the reasoning behind model predictions. Understanding the robustness of interpretation methods is important especially in sensitive domains such as medical applications since interpretation results are often used in downstream tasks. Although gradient-based saliency …
Survey on principles and challenges of interpretable machine learning.
problem Improving machine learning models' interpretability for high-stakes decisions.
method Identification and analysis of 10 technical challenges in interpretable machine learning.
result Identification of 10 technical challenges in interpretable machine learning.
Interpretable machine learning has become a strong competitor for traditional black-box models. However, the possible loss of the predictive performance for gaining interpretability is often inevitable, putting practitioners in a dilemma of choosing between high accuracy (black-box models) and interpretability (interpr…
The paper uses model-based trees to create interpretable surrogate models for complex machine learning models.
problem Interpreting complex machine learning models.
method Using model-based trees to partition feature space and create interpretable models.
result Model-based trees generate optimal surrogate models that balance interpretability and performance.
FF layers in transformers are nearly as interpretable as sparse autoencoders.
problem Comparing interpretability of feature vectors in FF layers vs. sparse autoencoders.
method Revisited interpretability of FF layers as key-value memories using modern benchmarks.
result FF and SAE feature vectors are similarly interpretable, but FFs can be better in some aspects.
Model interpretability is a requirement in many applications in which crucial decisions are made by users relying on a model's outputs. The recent movement for "algorithmic fairness" also stipulates explainability, and therefore interpretability of learning models. And yet the most successful contemporary Machine Learn…
Improving the interpretability of brain decoding approaches is of primary interest in many neuroimaging studies. Despite extensive studies of this type, at present, there is no formal definition for interpretability of brain decoding models. As a consequence, there is no quantitative measure for evaluating the interpre…
Framework evaluates post-hoc interpretability methods in time-series classification.
problem Lack of suitable post-hoc interpretability methods for time-series classification.
method Proposes a framework with quantitative metrics to assess interpretability methods.
result Addresses several drawbacks of existing methods, including dependence on human judgement and data distribution shift.
This paper explores causal analysis in machine learning for better interpretability.
problem The lack of causality in traditional interpretable machine learning models.
method An overview of causal approaches for interpretable machine learning.
result Causal analysis improves the interpretability of machine learning models.
When quantitative models are used to support decision-making on complex and important topics, understanding a model's ``reasoning'' can increase trust in its predictions, expose hidden biases, or reduce vulnerability to adversarial attacks. However, the concept of interpretability remains loosely defined and applicatio…
Most of the work on interpretable machine learning has focused on designing either inherently interpretable models, which typically trade-off accuracy for interpretability, or post-hoc explanation systems, which lack guarantees about their explanation quality. We propose an alternative to these approaches by directly r…