arXiv subjects

Abhijit Dasgupta

Publications and source records attributed to Abhijit Dasgupta.

3 recordsLinked to original sources

arXiv2026

Invocation-Level Reliability of Tool-Using Agents

Tool-using agents fail two ways: choosing the wrong tool, or forming wrong arguments, and an early failure of either kind can silently corrupt everything downstream. We measure a correct-invocation rate that separates the two, under both a clean teacher-forced context and the model's own free-running context, on five open-weight models over contamination-free multi-step tasks (depths 1-8). By depth 6, roughly 70% of a model's own clean-context capability is lost to its own earlier mistakes (L6 = 0.686, 0.684). Our central finding concerns the measurement itself. Under exact-match scoring against a fixed gold trajectory, a propagation model's severity and recovery parameters are not merely hard to estimate - they are fixed by the scoring rule. Severity is forced to its boundary (0 of 869 poisoned steps correct); recovery is structurally unobservable (0 of 580 poisoned steps returned on-track, against an expected 0.0058 by chance). Both follow from one mechanism: post-divergence, the gold value is generated by tool constants the model never sees, so it is information the model cannot derive. A fit run anyway returns 0.92 and 0.73 for a quantity that is exactly 1.000 - confident numbers for a parameter the scoring rule already determined. We give the mechanism and a remedy, conditional-on-state scoring, applied retrospectively to cached completions at zero additional cost, which un-pins severity to interior estimates excluding zero (+0.149, +0.316).

cs.AI↗

arXiv2024

Music Genre Classification: Ensemble Learning with Subcomponents-level Attention

Music Genre Classification is one of the most popular topics in the fields of Music Information Retrieval (MIR) and digital signal processing. Deep Learning has emerged as the top performer for classifying music genres among various methods. The letter introduces a novel approach by combining ensemble learning with attention to sub-components, aiming to enhance the accuracy of identifying music genres. The core innovation of our work is the proposal to classify the subcomponents of the music pieces separately, allowing our model to capture distinct characteristics from those sub components. By applying ensemble learning techniques to these individual classifications, we make the final classification decision on the genre of the music. The proposed method has superior advantages in terms of accuracy compared to the other state-of-the-art techniques trained and tested on the GTZAN dataset.

cs.SD↗

arXiv2024

Compactness and Symmetric Well Orders

We introduce and investigate a topological version of Stäckel's 1907 characterization of finite sets, with the goal of obtaining an interesting notion that characterizes usual compactness (or a close variant of it). Define a $T_2$ topological space $(X, τ)$ to be Stäckel-compact if there is some linear ordering $\prec$ on $X$ such that every non-empty $τ$-closed set contains a $\prec$-least and a $\prec$-greatest element. We find that compact spaces are Stäckel-compact but not conversely, and Stäckel-compact spaces are countably compact. The equivalence of Stäckel-compactness with countable compactness remains open, but our main result is that this equivalence holds in scattered spaces of Cantor-Bendixson rank $< ω_2$ under ZFC. Under V=L, the equivalence holds in all scattered spaces.

math.GN↗