arXiv subjects

Hokky Situngkir

Publications and source records attributed to Hokky Situngkir.

At least 19 recordsLinked to original sources

arXiv2026

Artificial Intelligence-Assisted Digital Inventory of Cultural Heritage & Traditional Knowledge: Case for Indonesian Open Digital Library of Culture

The Indonesian Digital Library of Culture (Perpustakaan Digital Budaya Indonesia, PDBI; budaya-indonesia.org) is a participatory platform that has collected tens of thousands of entries on Nusantara cultural heritage through public contribution since 2007. Manual contribution faces three structural barriers: coverage (knowledge is scattered across languages and sites), integrity (open sources mix authentic documentation with noise), and completeness (subjects are recorded but their data remain shallow). This paper presents a methodological framework for autonomous, AI-based harvesting of cultural knowledge from the open web, designed to expand corpus coverage while intensifying per-entry data depth. The methodology is organised as a five-stage economic funnel: focused crawling, multilingual extraction and canonicalisation, vector encoding with blocking, agentic decision-making, and idempotent publication, under the principle of deterministic orchestration, agentic decisions. Each stage is formalised: funnel economics and optimal filter ordering; crawl-frontier dynamics as a subcritical branching process that explains the necessity of recurrent re-seeding; fact-level novelty via a containment measure; Bayesian multi-source evidence fusion with elevated publication thresholds for sacred categories; exactly-once effects via idempotent upserts and the transactional outbox; sliding-window inference budgeting with a reservation protocol; statistical quality auditing; and seed selection as submodular coverage maximisation. The framework retains four high-value human roles: curator of direction, escalation approver, quality auditor, and guardian of meaning, while machine autonomy is raised in stages. Ethical, legal, and cultural-sensitivity implications are discussed, including the architectural guarantee that the machine never overwrites human contributions.

cs.AI↗

arXiv2026

Information Networks of Stock Prices

The collective movement of stock prices harbors complex interdependencies that are conventionally simplified only through a linear lens. This paper explores computed structural network representations in the Indonesian capital market by testing the limits of Pearson correlation and Mutual Information (MI) in unveiling the spectral dynamics of the market. Across 2,328 rolling observation windows from 2015 to 2025, we examine 24 methodological configurations that combine three dependency estimators (Pearson, MI adaptive binning, and MI-kNN), two graph filtering schemes (Minimum Spanning Tree/MST and Planar Maximally Filtered Graph/PMFG), and four community decoders. The empirical results unveil a fundamental reality: topological richness does not always resonate with sectoral classification precision. The Pearson, MST, and Infomap configuration is shown to remain the most robust foundation for recovering conventional sectoral taxonomy. Nevertheless, when deeper observation demands the exposition of local structures and the weave of heterogeneous communities, the architectural relaxation through PMFG demonstrates its superiority. In the realm of residual information detection, MI adaptive binning appears far more proportional than kNN; histogram-based regularization successfully tames empirical noise without sweeping away traces of non-linear dependency. Ultimately, the synergy of MI and PMFG is not positioned to dethrone the dominance of linear correlation, but rather to provide an essential analytical lens for excavating hidden economic sub-structures -- such as the cohesion of commodity regimes -- that have long transcended the rigid boundaries of the market's formal sectors.

cs.SI↗

arXiv2026

Arithmetic Pedagogy for Language Models

We investigate whether methods of human mathematics pedagogy can guide the training of language models toward arithmetic reasoning. Building on the GASING method -- an Indonesian pedagogy that solves basic arithmetic through a left-to-right procedure aligned with the causal order of token generation -- we operationalize each operation as a computational procedure whose execution trace is serialized into natural-language Chain-of-Thought (CoT) supervision. A small GPT-2 decoder (86M parameters) with a syllabic-agglutinative TOBA tokenizer for Indonesian is trained from scratch on this data using only a next-token prediction objective, without reinforcement learning or reward-based optimization. Monitoring training reveals three distinct learning phases, and mechanistic analyses -- attention-masking interventions on the CoT information graph, residual-stream probing, and logit-lens inspection -- show that the model first internalizes a procedural pathway and subsequently develops an associative, ``mental-arithmetic'' capacity that retrieves intermediate results without explicit step-by-step computation. The trained model reaches over 80% accuracy on held-out problems and attains competitive performance against substantially larger language models, indicating that targeted, pedagogically grounded training can yield strong and economical arithmetic capability at small scale.

cs.CL↗

arXiv2026

Adaptive Engram Memory System for Indonesian Language Model: Generative AI Based on TOBA LM for Batak and Minang Language

This study presents TOBA-LM, a trilingual language model based on GPT-2 architecture with 1.2 billion parameters, trained on a corpus encompassing Indonesian, Batak, and Minangkabau using syllabic-agglutinative tokenization. The architecture integrates an Engram Memory mechanism, an adaptive n-gram-based memory system with a 500,000 x 768 embedding table that captures morphological dependencies through bigram and trigram pathways. Empirical results demonstrate a training efficiency of 80%, with the loss value dropping from 6.4 to 1.7996 in only 12,973 steps -- significantly faster than the conventional transformer architecture, which required over 70,000 steps to achieve comparable convergence. These findings confirm that the integration of external statistical memory substantially reduces computational requirements for developing regional language models under limited resources.

cs.CL↗

arXiv2026

Tokenizations for Austronesian Language Models: study on languages in Indonesia Archipelago

Tokenization constitutes a fundamental stage in Large Language Model (LLM) processing; however, subword-based tokenization methods optimized on English-dominant corpora may produce token fragmentation misaligned with the linguistic structures of Austronesian languages. This study aimed to develop a syllable-based tokenization framework adopting principles from traditional Indonesian scripts (aksara) for regional languages of Indonesia. A syllabic segmentation procedure was constructed based on the logic of abugida writing systems and implemented with a vocabulary of 2,843 tokens extracted from the Indonesian dictionary (KBBI). Evaluation was conducted on the NusaX dataset comprising 1,000 parallel translation samples across 10 regional languages, Indonesian, and English. Analysis employed Token per Character (TPC) ratio and sequence alignment using the Smith-Waterman algorithm. Results demonstrated that syllable-based tokenization yielded consistent TPC values across all regional languages, whereas GPT-2 exhibited an inverse pattern with the lowest TPC for English. Syllable-based tokenization consistently produced higher token sequence similarity scores, with an average increase of approximately 21% compared to GPT-2. These findings confirm that the syllable-based approach more effectively preserves phonological and morphological patterns across related Austronesian languages, offering a linguistically principled foundation for multilingual LLM development.

cs.CY↗

arXiv2016

Agent-Based Model for River-Side Land-living: Portrait of Bandung Indonesian Cikapundung Park Case Study

A city park has been built from the organic urban settlement in the Cikapundung River, Bandung, Indonesia. While the aim for the development is the revitalization of the river for being unhealthy from the waste coming from the settlement. A study on how Indonesian people, in general, treating water source, like river, lake, and ocean is revisited. Throwing waste into the river has actually become paradox with the collective mental understanding about water among Indonesians. Two scenarios of agent-based simulation is presented, to see the dynamics of organic settlement and life of the city park after being opened for public. The simulation is delivered upon the imagery of landscape taken from the satellite and drone. While experience for presented problems gives insights, the computational social laboratory also awaits for further theoretical explorations and endeavors to sharpen good policymaking.

cs.CY↗

arXiv2015

On Capturing the Spreading Dynamics over Trading Prices in the Market

While market is a social field where information flows over the interacting agents, there have been not so many methods to observe the spreading information in the prices comprising the market. By incorporating the entropy transfer in information theory in its relation to the Granger causality, the paper proposes a tree of weighted directed graph of market to detect the changes of price might affect other price changes. We compare the proposed analysis with the similar tree representation built from the correlation coefficients of stock prices in order to have insight of possibility in seeing the collective behavior of the market in general.

q-fin.ST↗

arXiv2015

Deconstructing Bataknese Gorga Computationally

The carved and painted decorations in traditional Batak houses and buildings, gorga, are the source of their exoticism. There are no identical patterns of the ornaments within Batak houses and the drawings are closely related to the way ancient Batak capture the dynamicity of the growing 'tree of life', one of central things within their cosmology and mythology. The survey of ornaments of Batak houses and buildings in Northern Sumatera Indonesia has made us possible to observe the complex pattern. The fractal dimensions of the geometrical shapes in gorga are calculated and they are conjectured into 1.5-1.6, between the dimensional of a line and a plane. The way gorga is drawn is captured by using some modification to the turtle geometry of L-System model, a popular model to model the dynamics of growing plants. The result is a proposal to see Bataknese gorga as one of traditional heritage that may enrich the studies to the generative art.

cs.CG↗

arXiv2015

On Social and Economic Spheres: An Observation of the 'gantangan' Indonesian tradition

Indonesian traditional villagers have a tradition for the sake of their own social and economic security named 'nyumbang'. There are wide variations of the traditions across the archipelago, and we revisit an observation to one in Subang, West Java, Indonesia. The paper discusses and employs the evolutionary game theoretic insights to see the process of 'gantangan', of the intertwining social cohesion and economic expectation of the participation within the traditional activities. The current development of the gantangan tradition is approached and generalized to propose a view between the economic and social sphere surrounding modern people. While some explanations due to the current development of gantangan is drawn, some aspects related to traditional views complying the modern life with social and economic expectations is outlined.

cs.CY↗

arXiv2015

Exploring Ancient Architectural Designs with Cellular Automata

The paper discusses the utilization of three-dimensional cellular automata employing the two-dimensional totalistic cellular automata to simulate how simple rules could emerge a highly complex architectural designs of some Indonesian heritages. A detailed discussion is brought to see the simple rules applied in Borobudur Temple, the largest ancient Buddhist temple in the country with very complex detailed designs within. The simulation confirms some previous findings related to measurement of the temple as well as some other ancient buildings in Indonesia. This happens to open further exploitation of the explanatory power presented by cellular automata for complex architectural designs built by civilization not having any supporting sophisticated tools, even standard measurement systems.

cs.CY↗

arXiv2015

Borobudur was Built Algorithmically

The self-similarity of Indonesian Borobudur Temple is observed through the dimensionality of stupa that is hypothetically closely related to whole architectural body. Fractal dimension is calculated by using the cube counting method and found that the dimension is 2.325, which is laid between the two-dimensional plane and three dimensional space. The applied fractal geometry and self-similarity of the building is emerged as the building process implement the metric rules, since there is no universal metric standard known in ancient traditional Javanese culture thus the architecture is not based on final master plan. The paper also proposes how the hypothetical algorithmic architecture might be applied computationally in order to see some experimental generations of similar building. The paper ends with some conjectures for further challenge and insights related to fractal geometry in Javanese traditional cultural heritages.

cs.CY↗

arXiv2015

Cellular-Automata and Innovation within Indonesian Traditional Weaving Crafts: A Discourse of Human-Computer Interaction

The paper reports the possibility of Indonesian traditional artisans of weaving designs and crafts to explore the cellular automata, a dynamical model in computation that may yield similar patterns. The reviews of the cellular automata due to the perspective of weaving process reveals that the latter would focus on macro-properties, i.e.: the strength of structural construction beside the aesthetic patterns and designs. The meeting of traditional weaving practice and the computational model is delivered and open the door for interesting discourse of computer-aided designs for the traditional artists and designers to come.

cs.CY↗

arXiv2015

Indonesia embraces the Data Science

The information era is the time when information is not only largely generated, but also vastly processed in order to extract and generated more information. The complex nature of modern living is represented by the various kind of data. Data can be in the forms of signals, images, texts, or manifolds resembling the horizon of observation. The task of the emerging data sciences are to extract information from the data, for people gain new insights of the complex world. The insights may came from the new way of the data representation, be it a visualizations, mapping, or other. The insights may also come from the implementation of mathematical analysis and or computational processing giving new insights of what the states of the nature represented by the data. Both ways implement the methodologies reducing the dimensionality of the data. The relations between the two functions, representation and analysis are the heart of how information in data is transformed mathematically and computationally into new information. The paper discusses some practices, along with various data coming from the social life in Indonesia to gain new insights about Indonesia in the emerging data sciences. The data sciences in Indonesia has made Indonesian Data Cartograms, Indonesian Celebrity Sentiment Mapping, Ethno-Clustering Maps, social media community detection, and a lot more to come, become possible. All of these are depicted as the exemplifications on how Data Science has become integral part of the technology bringing data closer to people.

cs.CY↗

arXiv2005

What can we see from Investment Simulation based on Generalized (m,2)-Zipf law?

The paper revisits the investment simulation based on strategies exhibited by Generalized (m,2)-Zipf law to present an interesting characterization of the wildness in financial time series. The investigations of dominant strategies on each specific time series shows that longer words dominant in larger time scale exhibit shorter dominant ones in smaller time scale and vice versa. Moreover, denoting the term wildness based on persistence over short term trend and memory represented by particular length of words, we can see how wild historical fluctuations over time series data coped with the Zipf strategies.

physics.soc-ph↗

arXiv2005

How Far Can We Go Through Social System?: Algorithmic Information Social Theories

The paper elaborates an endeavor on applying the algorithmic information-theoretic computational complexity to meta-social-sciences. It is motivated by the effort on seeking the impact of the well-known incompleteness theorem to the scientific methodology approaching social phenomena. The paper uses the binary string as the model of social phenomena to gain understanding on some problems faced in the philosophy of social sciences or some traps in sociological theories. The paper ends on showing the great opportunity in recent social researches and some boundaries that limit them.

nlin.CD↗

arXiv2004

Innovation as Evolution: Case Study Phylomemetic of Cellphone Designs

Cellular phone is one of the most developing technological artifacts today. The evolution occurs through random innovation. Our effort is trying to view the evolution of this artifact from memetics. By constructing a phylomemetic tree based on cellular phone memes to infer or estimate the evolutionary history and relationship among cellular phone. We adopt several methods, which are commonly used in constructing phylogenetic tree, they are UPGMA algorithm and Parsimony Maximum algorithm to construct cellphone phylomemetic tree. Therefore we compare with the innovation tree, which is based on serial number and their appearance time. From phylomemetic tree, we then analyze the process of a cellular phone innovation through looking out on the cellular phone type lies in the same cluster. The comparison of the simulation tree result shows a generally different branching pattern, giving a presumption that innovation in cellular phone is not really relating with their serial number, but occurs merely because of random mutation of allomeme design and competes with its technological development.

nlin.AO↗

arXiv2004

Power Law Signature in Indonesian Population

The paper analyzes the spreading of population in Indonesia. The spreading of population in Indonesia is clustered in two regional terms, i.e.: kabupaten and kotamadya. It is interestingly found that the rank in all kabupaten respect to the population does not have fat tail properties, while in the other hand; there exists power-law signature in kotamadya. We analyzed that this fact could be caused by the equal or similar infrastructural development in all regions; nevertheless, we also note that the first 20 kabupatens are dominated in Java and Sumatera. Furthermore, the fat tail character in the rank of kotamadya could be caused by the big gap between big cities one another, e.g.: Jakarta, Surabaya, and others. The paper ends with some suggestions of more attention to infrastructural development in eastern regional cities.

nlin.PS↗

arXiv2004

Statistical Facts of Artificial Stock Market

The paper reports the construction of artificial stock market that emerges the similar statistical facts with real data in Indonesian stock market. We use the individual but dominant data, i.e.: PT TELKOM in hourly interval. The artificial stock market shows standard statistical facts, e.g.: volatility clustering, the excess kurtosis of the distribution of return, and the scaling properties with its breakdown in the crossover of Levy distribution to the Gaussian one. From this point, the artificial stock market will always be evaluated in order to have comprehension about market process in Indonesian stock market generally.

cond-mat.other↗