Services
Responsible AI
Is your machine learning model fair enough? What if the lending AI discriminates on gender and race? What if the accuracy of medical AI depends on a person’s annual income or on the GDP of the country where it is used? Today’s AI has the potential to cause such problems.
To prevent harmful outcomes, we research and identify best fitted libraries to develop Responsible AI metrics, a governance framework that promotes the practice of building generative AI that is reliable, transparent, accountable and ethical. We help our clients implement a safe and non-discriminatory model that consist of these factors:
-
Bias
How do you prevent your model from making predictions that do not favor or discriminate against certain individuals or groups. Bias can impact machine learning systems at pretty much every stage. Bias is a preference or prejudice against a particular group, individual, or feature and comes in many forms.
-
Drift
Generative AI models are trained on historical data to learn a static mapping between their input and output variables and operate within spec. Drift occurs when the models are deployed on continuously streamed data, whose nature is likely to change over time (data or concept drift), model performance may suddenly and substantially degrade, forcing continuously update the models to reflect the new data distribution.
-
Explainability
Being able to understand why the model generated a certain result and explain why certain prediction was made. It is about interrogating a model, gathering information on why a particular prediction (or series of predictions) was made, and understand the model and instance view. Instance View tells you for a particular prediction, what factors contributed.
-
Data Privacy
Sometimes the training and input data can be quite sensitive and it is essential to consider the potential privacy implications in using sensitive data. This includes not only respecting the legal and regulatory requirements, but also considering social norms and typical individual expectations.
-
Carbon Footprint
It estimates the amount of carbon dioxide (CO2) produced by the cloud or personal computing resources used to execute the AI code.
Exploratory Data Analysis (EDA)
Exploratory Data Analysis is the critical process of running initial investigations on data so as to discover patterns, to spot anomalies, to test hypothesis and to check assumptions. EDA is an analysis approach that identifies general patterns in the data. These patterns include outliers and features of the data that might be unexpected.
We have a deep bench of “data hunters” with extensive experience in time series data that help our clients unlock the hidden patterns.
Why
Why Us
Artificial intelligence and generative model development is not a new field. Rewind the clock to the 1990s, the golden age of financial engineering. This was the time that the derivatives market exploded, fueled by the ease of analysis and innovation to create new financial products using Microsoft Excel. The cold war ended and the mathematicians from the defense industry flocked to wall street to pioneer the field of electronic trading.
It was the era of “rise of the machines”. The time to create and maintain massive time series databases, analytics models and engines, stochastic calculus and Montecarlo simulation, performance feedback loops, etc.
This is our history. We worked at Credit Suisse, Merrill Lynch and Bear Stearns and developed Algorithmic trading software to review the market data and locate the dislocations in the market for profitable electronic trading.
Let us help you establish a correct trajectory for your generative AI development.
Contact
Contact Us
New York
175 Varick Street, 8th floor
New York, NY 10014