research | Ege Demirci

2025

ACL 2025 - SRW

Are LLMs Truly Graph-Savvy? A Comprehensive Evaluation of Graph Generation

Ege Demirci, Rithwik Kerur, and Ambuj Singh

Association for Computational Linguistics - ACL, Aug 2025

Abs PDF

While large language models (LLMs) have demonstrated impressive capabilities across diverse tasks, their ability to generate valid graph structures remains underexplored. We evaluate fifteen state-of-the-art LLMs on five specialized graph generation tasks spanning delivery networks, social networks, quantum circuits, gene-disease networks, and transportation systems. We also test the LLMs using 3 different prompt types: direct, iterative feedback, and program-augmented. Models supported with explicit reasoning modules (o3-mini-high, o1, Claude 3.7 Sonnet, DeepSeek-R1) solve more than twice as many tasks as their general-purpose peers, independent of parameter count. Error forensics reveals two recurring failure modes: smaller parameter size Llama models often violate basic structural constraints, whereas Claude models respect topology but mismanage higher-order logical rules. Allowing models to refine their answers iteratively yields uneven gains, underscoring fundamental differences in error-correction capacity. This work demonstrates that graph competence stems from specialized training methodologies rather than scale, establishing a framework for developing truly graph-savvy language models.
Preprint

From Occasional to Steady: Habit Formation Insights From a Comprehensive Fitness Study

Ege Demirci, Efe Tüzün, Ahmet Furkan Un, Taner Giray Sonmez, and Onur Varol

Under review, Jan 2025

Abs PDF

Exercising regularly is widely recognized as a cornerstone of health, yet the challenge of sustaining consistent exercise habits persists. Understanding the factors that influence the formation of these habits is crucial for developing effective interventions. This study utilizes data from Mars Athletic Club, Türkiye’s largest sports chain, to investigate the dynamics of gym attendance and habit formation. The general problem addressed by this study is identifying the critical periods and factors that contribute to the successful establishment of consistent exercise routines among gym-goers. Here we show that there are specific periods during which gym attendance is most crucial for habit formation. By developing a survival metric based on gym attendance patterns, we pinpoint these critical periods and segment members into distinct clusters based on their visit patterns. Our analysis reveals significant differences in how various subgroups respond to interventions, such as group classes, personal trainer sessions, and visiting different clubs. Using causal inference analysis, we demonstrate that personalized guidance and social dynamics are key drivers of sustained long-term engagement. By systematically examining these variables and considering the specific characteristics of different clusters, our research demonstrates the importance of a tailored, multi-dimensional approach to promoting exercise habits, which integrates social dynamics, personalized guidance, and strategic interventions to sustain long-term engagement.
Preprint

GroupGAT: Physics–Aware, Symmetry–Preserving Graph Attention for Network Flow Completion

Ege Demirci, Francesco Bullo, Ananthram Swami, and Ambuj Singh

Under review for NeurIPS, Aug 2025

Abs

Recovering missing flows on the edges of a network, while exactly respecting local conservation laws, is a fundamental inverse problem that arises in many systems such as transportation, energy, and mobility. We introduce GROUPGAT, a novel architecture that combines (i) a group-action on divergence-free flows, (ii) a graph-attention encoder to learn feature-conditioned weights over these symmetry- preserving actions, and (iii) a lightweight Tikhonov refinement solved via implicit bilevel optimization. The method first anchors the given observation on a minimum- norm divergence-free completion. We then compute an orthonormal basis for all admissible group actions that leave the observed flows invariant and endow the flow manifold with an Abelian group structure under vector addition. A stack of GATv2 layers then encodes the graph and its edge features into per-edge embeddings, which are pooled over the missing edges and produce per-basis attention weights. This attention-guided process selects a set of physics-aware group actions that preserve the observed flows. Finally, a scalar Tikhonov penalty refines the missing entries via a convex least-squares solver, with gradients propagated implicitly through Cholesky factorization. Across three real-world flow benchmarks (traffic, power, bike), GROUPGAT substantially outperforms state-of-the-art baselines in RMSE, MAE and correlation metrics.
Preprint

Graph Neural Networks Should Learn Symmetry: A Third-Wave Agenda for Automatic and Adaptive Equivariance

Ege Demirci, and Ambuj Singh

Under review for NeurIPS Position Paper Track, Aug 2025

Abs

This position paper argues that Graph Neural Networks (GNNs) should learn symmetry automatically and adaptively rather than rely on hand-coded or implicit invariances. The first wave of GNNs relied on implicit permutation invariance; the second wave hard-wired richer groups such as E(n) or SE(3), achieving state-of-the-art results yet failing when real data violate ideal symmetry. We outline a third wave built on (i) automatic discovery of latent symmetry gener- ators (LieConv, LieGAN, AtlasD); (ii) soft, layer-wise relaxation that learns how much symmetry to retain (Adaptive-E, REMUL, SymPE); (iii) dynamic-graph equivariance for evolving or heterogeneous structures (PEnGUiN, ESTAG); and (iv) task-equivariant meta-learning that transfers geometric priors across do- mains. Preliminary bounds show adaptive equivariance contracts hypothesis-class complexity and inherit universal-approximation properties of fixed-group models; closing the gap to full guarantees is an open research frontier. Across molecules, multi-agent control, and continuum-scale PDEs, third-wave architectures cut data or compute by an order of magnitude relative to both non-equivariant and rigidly equivariant baselines. We conclude with a seven-point agenda arguing that elevat- ing symmetry from a fixed prior to a learnable resource is key to greener models, safer controllers, and faster scientific discovery.

2024

Scientific Reports

First public dataset to study 2023 Turkish general election

Ali Najafi, Nihat Mugurtay, Yasser Zouzou, Ege Demirci, Serhat Demirkiran, Huseyin Alper Karadeniz, and Onur Varol

Scientific Reports, Mar 2024

Abs PDF

In the context of Turkiye’s most recent parliamentary and presidential elections (“seçim” in Turkish), social media has played an important role in shaping public debate. It is of utmost importance to capture social media trends during the 2023 Turkish elections, since it uncovers a great deal of information of election propaganda, political debates, smear campaigns, and election manipulation by domestic and international actors. We provide a comprehensive dataset for social media researchers to study Turkish elections, develop tools to prevent online manipulation, and gather novel information to inform the public. We are committed to continually improving the data collection and updating it regularly leading up to the election. Using the #Secim2023 dataset, researchers can examine the social and communication networks between political actors, track current trends, and investigate emerging threats to election integrity. Our dataset and analysis code available through Harvard Dataverse and Github, respectively.