اشترك بالحزمة الذهبية واحصل على وصول غير محدود شمرا أكاديميا
تسجيل مستخدم جديدA Holistic Approach to Interpretability in Financial Lending: Models, Visualizations, and Summary-Explanations
70
0
0.0
(
0
)
اسأل ChatGPT حول البحث
ﻻ يوجد ملخص باللغة العربية
Lending decisions are usually made with proprietary models that provide minimally acceptable explanations to users. In a future world without such secrecy, what decision support tools would one want to use for justified lending decisions? This question is timely, since the economy has dramatically shifted due to a pandemic, and a massive number of new loans will be necessary in the short term. We propose a framework for such decisions, including a globally interpretable machine learning model, an interactive visualization of it, and several types of summaries and explanations for any given decision. The machine learning model is a two-layer additive risk model, which resembles a two-layer neural network, but is decomposable into subscales. In this model, each node in the first (hidden) layer represents a meaningful subscale model, and all of the nonlinearities are transparent. Our online visualization tool allows exploration of this model, showing precisely how it came to its conclusion. We provide three types of explanations that are simpler than, but consistent with, the global model: case-based reasoning explanations that use neighboring past cases, a set of features that were the most important for the models prediction, and summary-explanations that provide a customized sparse explanation for any particular lending decision made by the model. Our framework earned the FICO recognition award for the Explainable Machine Learning Challenge, which was the first public challenge in the domain of explainable machine learning.
قيم البحث
اقرأ أيضاً
As machine learning and algorithmic decision making systems are increasingly being leveraged in high-stakes human-in-the-loop settings, there is a pressing need to understand the rationale of their predictions. Researchers have responded to this need
with explainable AI (XAI), but often proclaim interpretability axiomatically without evaluation. When these systems are evaluated, they are often tested through offline simulations with proxy metrics of interpretability (such as model complexity). We empirically evaluate the veracity of three common interpretability assumptions through a large scale human-subjects experiment with a simple placebo explanation control. We find that feature attribution explanations provide marginal utility in our task for a human decision maker and in certain cases result in worse decisions due to cognitive and contextual confounders. This result challenges the assumed universal benefit of applying these methods and we hope this work will underscore the importance of human evaluation in XAI research. Supplemental materials -- including anonymized data from the experiment, code to replicate the study, an interactive demo of the experiment, and the models used in the analysis -- can be found at: https://doi.pizza/challenging-xai.
We present the Language Interpretability Tool (LIT), an open-source platform for visualization and understanding of NLP models. We focus on core questions about model behavior: Why did my model make this prediction? When does it perform poorly? What
happens under a controlled change in the input? LIT integrates local explanations, aggregate analysis, and counterfactual generation into a streamlined, browser-based interface to enable rapid exploration and error analysis. We include case studies for a diverse set of workflows, including exploring counterfactuals for sentiment analysis, measuring gender bias in coreference systems, and exploring local behavior in text generation. LIT supports a wide range of models--including classification, seq2seq, and structured prediction--and is highly extensible through a declarative, framework-agnostic API. LIT is under active development, with code and full documentation available at https://github.com/pair-code/lit.
I prove that a centre manifold approach to creating finite difference models will consistently model linear dynamics as the grid spacing becomes small. Using such tools of dynamical systems theory gives new assurances about the quality of finite diff
erence models under nonlinear and other perturbations on grids with finite spacing. For example, the advection-diffusion equation is found to be stably modelled for all advection speeds and all grid spacing. The theorems establish an extremely good form for the artificial internal boundary conditions that need to be introduced to apply centre manifold theory. When numerically solving nonlinear partial differential equations, this approach can be used to derive systematically finite difference models which automatically have excellent characteristics. Their good performance for finite grid spacing implies that fewer grid points may be used and consequently there will be less difficulties with stiff rapidly decaying modes in continuum problems.
As machine learning algorithms getting adopted in an ever-increasing number of applications, interpretation has emerged as a crucial desideratum. In this paper, we propose a mathematical definition for the human-interpretable model. In particular, we
define interpretability between two information process systems. If a prediction model is interpretable by a human recognition system based on the above interpretability definition, the prediction model is defined as a completely human-interpretable model. We further design a practical framework to train a completely human-interpretable model by user interactions. Experiments on image datasets show the advantages of our proposed model in two aspects: 1) The completely human-interpretable model can provide an entire decision-making process that is human-understandable; 2) The completely human-interpretable model is more robust against adversarial attacks.
By considering the various CEMP subclasses separately, we try to derive, from the specific signatures imprinted on the abundances, parameters (such as metallicity, mass, temperature, and neutron source) characterizing AGB nucleosynthesis from the spe
cific signatures imprinted on the abundances, and separate them from the impact of thermohaline mixing, first dredge-up, and dilution associated with the mass transfer from the companion.To put CEMP stars in a broad context, we collect abundances for about 180 stars of various metallicities, luminosity classes, and abundance patterns, from our own sample and from literature. First, we show that there are CEMP stars which share the properties of CEMP-s stars and CEMP-no stars (which we call CEMP-low-s stars). We also show that there is a strong correlation between Ba and C abundances in the s-only CEMP stars. This strongly points at the operation of the 13C neutron source in low-mass AGB stars. For the CEMP-rs stars (seemingly enriched with elements from both the s- and r-processes), the correlation of the N abundances with abundances of heavy elements from the 2nd and 3rd s-process peaks bears instead the signature of the 22Ne neutron source. Adding the fact that CEMP-rs stars exhibit O and Mg enhancements, we conclude that extremely hot conditions prevailed during the thermal pulses of the contaminating AGB stars. Finally, we argue that most CEMP-no stars (with no overabundances for the neutron-capture elements) are likely the extremely metal-poor counterparts of CEMP neutron-capture-rich stars. We also show that the C enhancement in CEMP-no stars declines with metallicity at extremely low metallicity ([Fe/H]~< -3.2). This trend is not predicted by any of the current AGB models.
الأسئلة المقترحة
سجل دخول لتتمكن من نشر تعليقات
التعليقات
جاري جلب التعليقات
سجل دخول لتتمكن من متابعة معايير البحث التي قمت باختيارها
