Posts

AI, Cybersecurity, and Superforecasting: What the Next Year Holds

Artificial Intelligence

Cybersecurity

The rapid evolution of Artificial Intelligence (AI) presents both transformative opportunities and complex risks.
The 2024 CCAPAC Report on AI and Cybersecurity highlighted…

Artificial Intelligence and the Shifting Balance of Power

Artificial Intelligence

Artificial Intelligence (AI) is no longer confined to research labs or experimental pilot programs. Today, it is a decisive factor in business strategy, shaping how…

Intractable inference problems

Evidence Lower Bound

Intractable Inference

Sources: Matthew N. Bernstein blog, Deep Learning Book, Ian Good Fellow

Solving Inverse Problems with score based generative modeling

Solving Inverse Problems

An inverse problem seeks to recover unobserved causal factors from a set of observed measurements.If \(x \in \mathbb{R}^n\) is an unknown signal and \(y \in…

Publication:In silico Antibody-Peptide Epitope prediction for Personalized cancer therapy

Personalized therapy

networks

machine learning

The human leukocyte antigen (HLA) system is a complex of genes on chromosome 6 in humans that encodes cell-surface proteins responsible for regulating the immune system.…

Are tumors patient specific

Personalized therapy

networks

machine learning

deep learning

To address this question we compare the characteristics and derived network measures of antibody–peptide epitopes for patient 1 and patient 2 to show the patient-specific…

RA-DIT RETRIEVAL-AUGMENTED DUAL INSTRUCTION TUNING

RA-DIT

LLM

RAG

In this post, I explain the approach called Retrieval-Augmented Dual Instruction Tuning (RA-DIT) (Lin et al., n.d.) that leverages both language model (LM) fine-tuning and…

Policy gradient methods

Reinforcement Learning

Reinforce

Policy gradient methods

Policy approximations

Sources: (“Sutton & Barto Book: Reinforcement Learning: An Introduction,” n.d.; Weng 2018)