Small Language Models (SLMs) in Enterprise: A Focused Approach to AI

One size does not fit all. Large language models (LLMs) like GPT-4 have certainly grabbed headlines with their broad knowledge and versatility. Yet, there’s a growing sense that sometimes, bigger isn’t always better, especially for enterprise applications. In this post, I want to dive into the world of Small Language Models (SLMs) and discuss why they might be the better fit for businesses aiming to harness AI in a more focused and efficient way.

​

Artificial Intelligence: A Historical Journey Introduction (AI/ML/DL)

Exploring the Evolution of Artificial Intelligence: Milestones, Breakthroughs, and Future Prospects. The fascination with artificial intelligence (AI) dates back centuries, with ancient myths depicting mechanical beings created by gods. In the scientific arena, early visionaries like Ada Lovelace and Charles Babbage laid the groundwork with their ideas of programmable machines.

​

Attention Mechanism

Introduction to Attention Mechanism with example. Covering the self-attention mechanism, the idea of query, key, and value, and discussing the multi-head attention.

​

Unlocking the Power of Clustering: A Beginner’s Guide

Clustering is an unsupervised machine learning technique that involves dividing a set of unlabeled samples into groups, or clusters, based on their similarity. Clustering can be used for exploring data, finding anomalies, and extracting features. It can be challenging to know how many groups to create.

​

​

Temporal Convolutional Networks, The Next Revolution for Time-Series

This post reviews the latest innovations of TCN-based solutions. We first present a case study of motion detection and briefly review the TCN architecture and its advantages over conventional approaches such as Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN). Then, we introduce several novels using TCN, including improving traffic prediction, sound event localization & detection, and probabilistic forecasting.

​

Is F1 the appropriate criterion to use? What about F2, F3,…, F beta?

It is very common to use the F1 measure for binary classification. This is known as the Harmonic Mean. However, a more generic F-beta score criterion might better evaluate model performance. So, what about F2, F3, and F-beta? 

 

Graph Convolutional Network for Time Series — An Intro

Graph convolutional network (GCN) is an absolute game-changer in the deep learning domain.

​

Learning Transfer Learning

Transfer learning is the process of using skills and knowledge, that have been learned in one situation to solve a different, related problem. Transfer learning is often viewed as a design methodology, as it involves applying previously learned information to new situations to improve the efficiency and effectiveness of the learning process. In other words, transfer learning allows individuals or machine learning algorithms to build upon their existing knowledge and skills to solve new problems. 

​

Solving The Class Imbalance Problem

Class imbalance is a common issue where the distribution of examples within a dataset is skewed or biased. This occurs when the training dataset has an unequal distribution of classes, leading to a potential bias in the trained model. Examples of imbalanced classification problems include fraud detection, claim prediction, default prediction, churn prediction, spam detection, anomaly detection, and outlier detection.

​

Pushing Explainable AI: Neural Networks Are Decision Trees

Exploring a new paper that aims to explain DNN behaviors.

​

Online Deep Learning (ODL) and Hedge Back-Propagation

Online learning is an ML method in which data is available in sequential order, and we use it in order to predict future data at each time step. Online Deep Learning is very challenging as it cannot use back-propagation.

​

Value-based Methods in Deep Reinforcement Learning

Deep Reinforcement learning has been a rising field in the last few years. A good approach to start with is the value-based method, where the state (or state-action) values are learned. In this post, a comprehensive review is provided where we focus on Q-learning and its extensions.

​

Tuning Q matrix for CV and CA models in Kalman Filter

Kalman Filter (KF) is widely used for vehicle navigation tasks, and in particular for vehicle trajectory smoothing. One of the problems associated while applying the KF for navigation tasks is the modeling of the vehicle trajectory.

​

Exploring The Latest Trends of Random Forest

The random forest model is considered one of the promising ML ensemble models that recently became highly popular. In this post, we review the last trends of the random forest.

​

Kalman Filter Celebrates 60 Years — An Intro.

The Kalman filter is one of the most influential ideas used in Engineering, Economics, and Computer Science for real-time applications. This year we mention 60 years for the novel publication.

​

AI-Based Worldwide Trends Due to COVID-19

COVID-19 has affected the worldwide economy, politics, education, tourism, and actually EVERYTHING. Many academic papers address trend prediction in various fields due to COVID-19, with the power of Artificial Intelligence.

​

Deep Learning in Geometry: Arclength Learning

A fundamental problem in geometry was solved using a Deep Neural Network (DNN). We learned a geometric property from examples in the supervised learning approach. As the simplest geometric object is a curve, we focused on learning the length of planar curves. For this reason, the fundamental length axioms were reconstructed and the ArcLengthNet was established.

​

The Exploding and Vanishing Gradients Problem in Time Series

In this post, we deal with exploding and Vanishing Gradient in Time Series and in particular in Recurrent Neural Networks (RNN) by Truncated BackPropagation Through Time and Gradient Clipping.

​

Penalizing the Discount Factor in Reinforcement Learning

The reinforcement learning field is used in many robotics problems and has a unique mechanism, where rewards should be accumulated through actions. But, what about the time between these actions?

​

Navigating in Tunnels Using a Smartphone Without Beacons or GPS

One of the main problems of autonomous vehicles is navigating in a GPS-denied environment. In this post, we focus on car positioning inside tunnels and present state-of-the-art accuracy in inertial navigation. The novel approach was developed by ALMA engineers combining deep learning, inertial sensors, and classical signal processing methods.

​

Deep Learning for Inertial Navigation

A short review of cutting-edge deep learning-based solutions for inertial navigation.

​

Deep Learning-based Inertial Navigation: A Hybrid Navigation Filter

The inertial navigation domain is considered a very classical one. It aims to provide us with navigation solutions (position, velocity, and orientation) using (low-graded) inertial sensors and accurate (low frequency) sensors, like the GPS receiver. A question has arisen about how powerful Deep Learning tools can boost this classical domain.

​

A World Without GPS: Can You Navigate Your Car with AI?

Imagine you wake up in the morning and check your navigation app, only to discover that “it is searching for a network”. Imagine that while you are driving, the navigation app suddenly stops working. Can you still navigate?

​

​