ADAbhiyan Dhakal
Menu

Questions, methods, evidence

Research

I am interested in making computing more efficient and predictable under resource constraints. Controlled, reproducible experiments are the method; resource behavior is the subject.

Current work

Current systems research on Linux memory pressure, responsiveness, and the signals available before a machine becomes unusable under load.

ongoing2026-present

Linux memory pressure and PSI-guided compression

Broader goal

How can memory-management policy preserve responsiveness when RAM is constrained?

Evidence to date

Controlled reruns show PSI before material swap growth, and a corrected canary records pressure-related responsiveness degradation. A three-host V5 study found that cgroup-memory growth rate identifies the tested moderate pressure regime about five seconds before PSI-only qualification and falls back cleanly under gradual pressure; earlier activation did not consistently outperform static zram.

Completed study and ongoing research

  1. Completed three-host guided-zram screens across structured/random workloads and abrupt-to-gradual pressure onset, followed by a focused confirmation with validated zram telemetry.

  2. Use sustained background-cgroup PSI with cgroup-memory growth rate as an early classifier; record PSI-full, swap I/O, and activation timing as diagnostic telemetry.

  3. Phase-align response/service tails, misses, occupancy, compression, writeback, reclaim, faults, and swap-path observations for each policy comparison.

Status and next stage

V1–V5 established early PSI/swap ordering, a valid responsiveness measure, compression sensitivity, and a reliable rate classifier with an inconclusive performance result. Research is ongoing. The evidence does not establish one globally beneficial guided-zram policy: static zram was best in the focused moderate/structured comparison, gradual/random outcomes were host-dependent, and activation timing alone is insufficient to explain responsiveness.

Previous applied research

Nepali legal information retrieval in a low-resource language setting.

The 2025 and 2026 ICAIL papers form a continuing empirical track: first examining retrieval-augmented legal analysis in Nepal, then isolating and comparing embedding models as a retrieval decision. This work provided experience with imperfect data, evaluation design, publication, and presenting results.

Publications

2026
preprint
arXiv:2608.13689
Preprint

A Bounded Reclaim Actuator for PSI-Guided Compressed Memory: A Controlled Ablation

Abhiyan Dhakal, Sanjog Sigdel

A controlled ablation of bounded reclaim for PSI-guided compressed memory under constrained physical memory.

Study contribution: Evaluates pressure-guided memory-management policies across foreground workloads, with explicit attention to responsiveness and workload-dependent effects.

PaperDataset
2026
conference
ICAIL 2026
Short paper; presented June 2026; not yet published

An Empirical Comparison of Embedding Models for Nepali Legal Document Retrieval

Abhiyan Dhakal, Pranish Kafle, Kausik Paudel, Sugat Sujakhu, Anita Jadhari, Prakash Poudyal

An empirical study of embedding choices for retrieving Nepali legal documents in a low-resource setting.

Study contribution: Compares retrieval behavior rather than treating embedding choice as an implementation detail, with attention to the constraints of Nepali legal text.

Conference program
2025
conference
National Conference on Computer Innovations (NCCI)
Published at NCCI 2025

An Artificial Intelligence Driven Semantic Similarity-Based Pipeline for Rapid Literature

Abhiyan Dhakal, Kausik Paudel, Sanjog Sigdel

A lightweight pipeline that generates search terms, retrieves open-access papers, ranks them with transformer embeddings, and applies statistical thresholding.

Study contribution: Evaluates three embedding models and documents a practical preliminary-review workflow without claiming ground-truth relevance labels that were not collected.

Paper
2025
demonstration
Proceedings of ICAIL 2025, pp. 498-499
Published · demonstration

Feasibility of Artificial Intelligence Driven Analysis in the Context of Nepalese Legal System

Abhiyan Dhakal, Sugat Sujakhu, Pranish Kafle, Kausik Paudel, Prakash Poudyal

A retrieval-augmented legal analysis pipeline designed for access to information in the Nepalese legal context.

Study contribution: Processes legal documents for retrieval, expands user queries, and evaluates rule-recall, rhetorical-understanding, and interpretation-oriented questions.

Paper