Abstract

Accurate multi-organ segmentation in abdominal CT is essential for diagnosis, treatment planning, and disease monitoring. However, it suffers from inaccurate segmentation due to intricate spatial relationships and varying organ shapes. Existing deep learning methods often struggle with imbalanced classes, size bias, and ambiguous boundaries, hindering segmentation accuracy. To address these challenges, this paper proposes Guided-nnUNet, a two-stage segmentation framework that decomposes abdominal multi-organ segmentation into organ localization and then localization-guided fine segmentation. In the first stage, a ResNet-50 model generates a low-dimensional localization map guiding organ locations. This spatial guidance is then fed into the second stage, where a 3D U-Net with dynamic affine feature-map transform performs the fine-grained segmentation by integrating spatial context from the localization map. Our evaluation on the publicly available AMOS and BTCV datasets demonstrates the effectiveness of the model. Guided-nnUNet achieves an average improvement of 7% and 9% on the AMOS and BTCV datasets, respectively, compared to the baseline nnUNet model. Additionally, our model outperforms the state-of-the-art MedNeXt by 3.6% and 5.3% on the AMOS and BTCV datasets, respectively. These results suggest that our two-stage solution offers a promising approach for accurate abdominal organ segmentation, particularly for overcoming the challenges associated with complex organ structures.

Abstract

We present a novel solution for accurate segmentation of pneumothorax from chest radiographs utilizing free-text radiology reports. Our solution employs text-guided attention to leverage the findings in the report to initially produce a low-dimensional region-localization map. These prior region maps are integrated at multiple scales in an encoder-decoder segmentation framework via dynamic affine feature map transform (DAFT). Extensive experiments on a public dataset CANDID-PTX, show that the integration of free-text reports significantly reduces the false positive predictions, while the DAFT-based fusion of localization maps improves the positive cases. Our method achieves a DSC of 0.60 for positive and 0.052 FPR for positive and negative cases, respectively, and 0.70 to 0.85 DSC for medium and large pneumothoraces.

Abstract

No Results Found

Abstract

We study a family of subcodes of the $m$-dimensional product code $mathscr{C}^{otimes m}$ (`subproduct codes') that have a recursive Plotkin-like structure, and which include Reed-Muller (RM) codes and Dual Berman codes as special cases. We denote the codes in this family as $mathscr{C}^{otimes [r,m]}$, where mbox{$0 leq r leq m$} is the `order' of the code. These codes allow a `projection' operation that can be exploited in iterative decoding, viz., the sum of two carefully chosen subvectors of any codeword in $mathscr{C}^{otimes [r,m]}$ belongs to $mathscr{C}^{otimes [r-1,m-1]}$. Recursive subproduct codes provide a wide range of rates and block lengths compared to RM codes while possessing several of their structural properties, such as the Plotkin-like design, the projection property, and fast ML decoding of first-order codes. Our simulation results for first-order and second-order codes, that are based on a belief propagation decoder and a local graph search algorithm, show instances of subproduct codes that perform either better than or within $0.5$~dB of comparable RM codes and CRC-aided Polar codes.

Abstract

Conversation is the most natural form of human communication, where each utterance can range over a variety of possible emotions. While significant work has been done towards the detection of emotions in text, relatively little work has been done towards finding the cause of the said emotions, especially in multimodal settings. SemEval 2024 introduces the task of Multimodal Emotion Cause Analysis in Conversations, which aims to extract emotions reflected in individual utterances in a conversation involving multiple modalities (textual, audio, and visual modalities) along with the corresponding utterances that were the cause for the emotion. In this paper, we propose models that tackle this task as an utterance labeling and a sequence labeling problem and perform a comparative study of these models, involving baselines using different encoders, using BiLSTM for adding contextual information of the conversation, and finally adding a CRF layer to try to model the inter-dependencies between adjacent utterances more effectively. In the official leaderboard for the task, our architecture was ranked 8th, achieving an F1-score of 0.1759 on the leaderboard.

Abstract

While significant work has been done in the field of NLP on vertical thinking, which involves primarily logical thinking, little work has been done towards lateral thinking, which involves looking at problems from an unconventional perspective and defying existing conceptions and notions. Towards this direction, SemEval 2024 introduces the task of BRAINTEASER, which involves two types of questions -- Sentence Puzzles and Word Puzzles that defy conventional common-sense reasoning and constraints. In this paper, we tackle both types of questions using few-shot prompting on GPT-3.5 and gain insights regarding the difference in the nature of the two types. Our prompting strategy placed us 26th on the leaderboard for the Sentence Puzzle and 15th on the Word Puzzle task.

Abstract

Large Language Models (LLMs) have showcased impressive abilities in generating fluent responses to diverse user queries. However, concerns regarding the potential misuse of such texts in journalism, educational, and academic contexts have surfaced. SemEval 2024 introduces the task of Multigenerator, Multidomain, and Multilingual Black-Box MachineGenerated Text Detection, aiming to develop automated systems for identifying machinegenerated text and detecting potential misuse. In this paper, we i) propose a RoBERTaBiLSTM based classifier designed to classify text into two categories: AI-generated or human ii) conduct a comparative study of our model with baseline approaches to evaluate its effectiveness. This paper contributes to the advancement of automatic text detection systems in addressing the challenges posed by machinegenerated text misuse. In the official leaderboard for the task, our architecture was ranked 46th, achieving an accuracy of 0.8083 on the leaderboard.

Abstract

Statistical analysis of high-frequency stock market order transaction data is conducted to understand order transition dynamics. We employ a first-order time-homogeneous discrete-time Markov chain model to the sequence of orders of stocks belonging to six different sectors during the US–China trade war of 2018. The Markov property of the order sequence is validated by the Chi-square test. We estimate the transition probability matrix of the sequence using maximum likelihood estimation. From the heatmap of these matrices, we found the presence of active participation by different types of traders during high volatility days. On such days, these traders place limit orders primarily with the intention of deleting the majority of them to influence the market. These findings are supported by high stationary distribution and low mean recurrence values of add and delete orders. Further, we found similar spectral gap and entropy rate values, which indicates that similar trading strategies are employed on both high and low volatility days during the trade war. Among all the sectors considered in this study, we observe that there is a recurring pattern of full execution orders in the Finance & Banking sector. This shows that the banking stocks are resilient during the trade war. Hence, this study may be useful in understanding stock market order dynamics and devise trading strategies accordingly on high and low volatility days during extreme macroeconomic events.

Abstract

The concept of sum labelling was introduced in 1990 by Harary. A graph is a sum graph if its vertices can be labelled by distinct positive integers in such a way that two vertices are connected by an edge if and only if the sum of their labels is the label of another vertex in the graph. It is easy to see that every sum graph has at least one isolated vertex, and every graph can be made a sum graph by adding at most n2 isolated vertices to it. The minimum number of isolated vertices that need to be added to a graph to make it a sum graph is called the sum number of the graph. The sum number of several prominent graph classes (e.g., cycles, trees, complete graphs) is already well known. We examine the effect of taking the disjoint union of graphs on the sum number. In particular, we provide a complete characterisation of the sum number of graphs of maximum degree two, since every such graph is the disjoint union of paths and cycles

Abstract

In this paper, we determine the computational complexity of recognizing two graph classes, grounded 𝖫-graphs and stabbable grid intersection graphs. An 𝖫-shape is made by joining the bottom end-point of a vertical (|) segment to the left end-point of a horizontal (−) segment. The top end- point of the vertical segment is known as the anchor of the 𝖫-shape. Grounded 𝖫-graphs are the intersection graphs of 𝖫-shapes such that all the 𝖫-shapes’ anchors lie on the same horizontal line. We show that recognizing grounded 𝖫-graphs is 𝖭𝖯-complete. This answers an open question asked by Jelínek & Töpfer (Electron. J. Comb., 2019). Grid intersection graphs are the intersection graphs of axis-parallel line segments in which two vertical (similarly, two horizontal) segments cannot intersect. We say that a (not necessarily axis-parallel) straight line 𝓁 stabs a segment 𝑠, if 𝑠 intersects 𝓁. A graph 𝐺 is a stabbable grid intersection graph (S TAB GIG) if there is a grid intersection representation of 𝐺 in which the same line stabs all its segments. We show that recognizing S TAB GIG graphs is 𝖭𝖯-complete, even on a restricted class of graphs. This answers an open question asked by Chaplick et al. (Order, 2018).