March 2026

Hi everyone, I’m working on a research project where my advisor suggested combining reinforcement learning with a transformer model, and I’m trying to figure out what the best architecture might look like. I unfortunately can’t share too many details about the actual project (sorry!), but I’ll try to explain the technical structure as clearly as possible using simplified examples. Problem setup (simplified example) Imagine we have a sequence where each element is represented by a super-token containing many […]

Ver mais

Table of Links Abstract and 1. Introduction 2. Pairing functions Enumerating trees LZ-trees Conclusion and References Abstract I present a simple algorithm for enumerating the trees generated by a Context Free Grammar (CFG). The algorithm uses a pairing function to form a bijection between CFG derivations and natural numbers, so that trees can be uniquely decoded from counting. This provides a general way to number expressions in natural logical languages, and potentially can be extended to other combinatorial […]

Ver mais

Textual Emotion Classification (TEC) is one of the most difficult NLP tasks. State of the art approaches rely on Large language models (LLMs) and multi-model ensembles. In this study, we challenge the assumption that larger scale or more complex models are necessary for improved performance. In order to improve logical consistency, We introduce CMHL, a novel single-model architecture that explicitly models the logical structure of emotions through three key innovations: (1) multi-task learning that jointly predicts primary emotions, […]

Ver mais

The classification of mental health is challenging for a variety of reasons. For one, there is overlap between the mental health issues. In addition, the signs of mental health issues depend on the context of the situation, making classification difficult. Although fine-tuning transformers has improved the performance for mental health classification, standard cross-entropy training tends to create entangled feature spaces and fails to utilize all the information the transformers contain. We present a new framework that focuses on […]

Ver mais

GitHub’s slopocalypse – the flood of AI-generated spam PRs and issues – has made Jazzband’s model of open membership and shared push access untenable. Jazzband was designed for a world where the worst case was someone accidentally merging the wrong PR. In a world where only 1 in 10 AI-generated PRs meets project standards, where curl had to shut down its bug bounty because confirmation rates dropped below 5%, and where GitHub’s own response was a kill switch to disable pull requests entirely – an […]

Ver mais

I was a speaker last month at the Pragmatic Summit in San Francisco, where I participated in a fireside chat session about Agentic Engineering hosted by Eric Lui from Statsig. The video is available on YouTube. Here are my highlights from the conversation. Stages of AI adoption We started by talking about the different phases a software developer goes through in adopting AI coding tools. 02:45 I feel like there are different stages of AI adoption as a […]

Ver mais

Modern Translation Systems heavily rely on high-quality, large parallel datasets for state-of-the-art performance. However, such resources are largely unavailable for most of the South Asian languages. Among them, Nepali and Tamang fall into such category, with Tamang being among the least digitally resourced languages in the region. This work addresses the gap by developing NepTam20K, a 20K gold standard parallel corpus, and NepTam80K, an 80K synthetic Nepali-Tamang parallel corpus, both sentence-aligned and designed to support machine translation. The […]

Ver mais

We entered the Release Candidate stage in the Godot 4.3 development cycle a week ago with 4.3 RC 1, which means that all features are in place, the most critical regressions have been tackled, and we’re confident that it’s now ready for general use in the vast majority of cases. A lot of users have been testing the RC 1 snapshot, reporting issues, many of which could be fixed. In the meantime, we’ve kept working on a few […]

Ver mais

We introduce Aumann-SHAP, an interaction-aware framework that decomposes counterfactual transitions by restricting the model to a local hypercube connecting baseline and counterfactual features. Each hyper-cube is decomposed into a grid in order to construct an induced micro-player cooperative game in which elementary grid-step moves become players. Shapley and LES values on this TU-micro-game yield: (i) within-pot contribution of each feature to the interaction with other features (interaction explainability), and (ii) the contribution of each instance and each feature […]

Ver mais