arXiv subjects

Halil Bisgin

Publications and source records attributed to Halil Bisgin.

3 recordsLinked to original sources

arXiv2026

Circumventing Magnitude Collapse via Modular Two-Stage Open-Set Recognition

Open-set recognition (OSR) requires neural networks to classify known classes while rejecting unknown samples, which is critical for real-world deployment. So far, OSR research studied and developed representation learning and postprocessing methods independently and their interactions remain unexplored. In this paper, we reconsider OSR as a modular two-stage pipeline and present the first systematic study of interactions between representation learning and postprocessing methods across dataset scales and usage of known unknown data. First, we discover a failure mode we term \emph{magnitude collapse}, where representation learning methods that utilize known unknown data can suffer performance degradation at large scale and irreversibly destroy discriminative information, despite excelling at small scale. Second, we study the modular interactions and reveal when they can be leveraged for modular performance gains via two-stage processing and where interaction effects amplify performance degradation due to magnitude collapse. Third, we show how these findings can be used to achieve state-of-the-art performance with a simple $K+1$ class classifier baseline and magnitude-aware postprocessing while circumventing magnitude collapse. Finally, our results demonstrate that small-scale evaluations with known unknown data are not predictive of large-scale performance, invalidating current best practices in OSR research. Code available at https://github.com/TwoDigitsOneNumber/Two-Stage-OSR.

cs.CV↗

arXiv2024

Large-Scale Evaluation of Open-Set Image Classification Techniques

The goal for classification is to correctly assign labels to unseen samples. However, most methods misclassify samples with unseen labels and assign them to one of the known classes. Open-Set Classification (OSC) algorithms aim to maximize both closed and open-set recognition capabilities. Recent studies showed the utility of such algorithms on small-scale data sets, but limited experimentation makes it difficult to assess their performances in real-world problems. Here, we provide a comprehensive comparison of various OSC algorithms, including training-based (SoftMax, Garbage, EOS) and post-processing methods (Maximum SoftMax Scores, Maximum Logit Scores, OpenMax, EVM, PROSER), the latter are applied on features from the former. We perform our evaluation on three large-scale protocols that mimic real-world challenges, where we train on known and negative open-set samples, and test on known and unknown instances. Our results show that EOS helps to improve performance of almost all post-processing algorithms. Particularly, OpenMax and PROSER are able to exploit better-trained networks, demonstrating the utility of hybrid models. However, while most algorithms work well on negative test samples -- samples of open-set classes seen during training -- they tend to perform poorly when tested on samples of previously unseen unknown classes, especially in challenging conditions.

cs.CV↗

arXiv2016

Attribution of Responsibility and Blame Regarding a Man-made Disaster: #FlintWaterCrisis

Attribution of responsibility and blame are important topics in political science especially as individuals tend to think of political issues in terms of questions of responsibility, and as blame carries far more weight in voting behavior than that of credit. However, surprisingly, there is a paucity of studies on the attribution of responsibility and blame in the field of disaster research. The Flint water crisis is a story of government failure at all levels. By studying microblog posts about it, we understand how citizens assign responsibility and blame regarding such a man-made disaster online. We form hypotheses based on social scientific theories in disaster research and then operationalize them on unobtrusive, observational social media data. In particular, we investigate the following phenomena: the source for blame; the partisan predisposition; the concerned geographies; and the contagion of complaining. This paper adds to the sociology of disasters research by exploiting a new, rarely used data source (the social web), and by employing new computational methods (such as sentiment analysis and retrospective cohort study design) on this new form of data. In this regard, this work should be seen as the first step toward drawing more challenging inferences on the sociology of disasters from "big social data".

cs.SI↗