论文检索

Thanh-Danh Nguyen, Vinh-Tiep Nguyen, Tam V. Nguyen

High-accuracy image segmentation models require abundant training annotated data which is costly for pixel-level annotations. Our work addresses a high-cost manual annotating process or the lack of detailed annotations via a generative approach. In particular, our approach (1) proposes the conditional instance-level synthesis to enrich the limited data to enhance the segmentation performance, and (2) employs the generative architectures to complete the segmentation task under few-shot learning concepts. The initial results on the Cityscapes benchmark emphasize our potential generative solution on the instance segmentation task given limited data.

Thye Shan Ng, Feiqi Cao, Soyeon Caren Han

Esports has rapidly emerged as a global phenomenon with an ever-expanding audience on livestream platforms. However, due to the complex nature of the game, it becomes challenging for newcomers to comprehend the gaming situation. This research introduces a 3M-Game that integrates multi-modal (MM) information from the livestream platform, including chat and livestream, to uncover the event. While conventional MM models typically prioritise aligning MM data through concurrent training towards a unified objective, our framework leverages multiple independent teachers trained on different tasks to accomplish game event detection. The results show the effectiveness of the proposed framework. The code and appendix are in https://github.com/adlnlp/3m_game.

Tzeh Yuan Neoh, Nicholas Teh

We study a model of sequential decision-making where voters have dynamic preferences over a set of candidates that are undesirable. This models scenarios such as the implementation of projects that are overall beneficial to society, but impose individual costs on certain affected individuals. We show that while minimizing the sum of agents' disutilities can be done in polynomial time, minimizing the maximum disutility obtained by any agent is computationally intractable, even in restricted cases. We then examine the potential for agents to engage in strategic manipulation in response to these welfare objectives, offering insights into possible misconduct within such decision-making environments.

Sujoy Nath, Arkaprabha Basu, Kushal Bose, Swagatam Das

In the rapidly advancing field of AI-assisted medical diagnosis, the generation of medical reports for Chest X-rays (CXR) has significantly improved with the increased availability of radiographs and their corresponding reports. However, these reports often contain complex medical terminology, making them difficult for patients and non-healthcare professionals to understand. In this study, we introduce a strategy called Chained Prompting for Improved Readability of Medical Reports (CPIR-MR), which translates original medical reports into more comprehensible language. Our primary contribution is the creation of a new extension to the IU X-Ray dataset, providing Simplified Medical Reports (SMRs) generated by CPIR-MR. Additionally, we demonstrate that standard methodologies can effectively produce these simplified reports by proposing a multi-modal text decoder (MTD) that combines BLIP with a classification network to generate simplified medical explanations (SMEs) when fine-tuned on SMRs.

Omkar Narkar, Aman Vohra, Ashiqur R. KhudaBukhsh

This study investigates user engagement and political polarization on YouTube Shorts, a special category of YouTube videos with a duration of 15-60 seconds. Via a substantial corpus of 38,838 videos gleaned from 100 YouTube channels focusing on political content, we contrast YouTube Shorts with long-form content in terms of user engagement, content toxicity, and polarization. Our analyses reveal that (1) YouTube Shorts receive more likes and views and fewer comments as compared to their long-form video counterparts; (2) YouTube Shorts are more toxic; and (3) considerably more polarized than long-form YouTube videos.

Siddharth Narendra, Shubham Ojha, Aditya Narendra, Abhay Kshirsagar, Abhisek Mallick

Conformal Prediction (CP) is an uncertainty quantification framework that provides prediction sets with a user-specified probability to include the true class in the prediction set. This guarantee on the user-specified probability is known as marginal coverage. Marginal coverage refers to the probability that the true label is included in the prediction set, averaged over all test samples. However, this can lead to inconsistent coverage across different classes, constraining its suitability for high-stakes applications such as pathological workflows. This study implements a Classwise CP method applied to two cancer datasets to achieve class-conditional coverage which ensures that each class has a user-specified probability of being included in the prediction set when it is the true label. Our results demonstrate the effectiveness of this approach through a significant reduction in the average class coverage gap compared to the Baseline CP method.

Itsuki Nakayama, Makoto Onizuka

Recently, researchers have focused on methods that not only distill knowledge from a Graph Neural Network (GNN) into a Multi-Layer Perceptron (MLP) but also leverage multiple teacher GNNs. However, existing methods assign a single attention weight to each teacher GNN. We propose a NodeAware Attention Mechanism (NAAM) that flexibly adjusts the attention weight for each node to leverage multiple GNNs fully. Experimental results show that NAAM outperforms existing GNN-to-MLP methods. our source code is available at: https://github.com/NakayamaItsuki/NAAM.

Atharv Mittal

Due to increasing privacy regulations and regulatory compliance, Machine Unlearning (MU) has become essential. The goal of unlearning is to remove information related to a specific class from a model. Traditional approaches achieve exact unlearning by retraining the model on the remaining dataset, but incur high computational costs. This has driven the development of more efficient unlearning techniques, including model sparsification techniques, which boost computational efficiency, but degrade the model’s performance on the remaining classes. To mitigate these issues, we propose a novel method, PruneLoRA which introduces a new MU paradigm, termed prune first, then adapt, then unlearn. LoRA reduces the need for large-scale parameter updates by applying low-rank updates to the model. We leverage LoRA to selectively modify a subset of the pruned model’s parameters, thereby reducing the computational cost, memory requirements and improving the model’s ability to retain performance on the remaining classes. Experimental Results across various metrics showcase that our method outperforms other approximate MU methods and bridges the gap between exact and approximate unlearning. Our code is available at https://github.com/vlgiitr/LoRA-Unlearn.

Ida Mattsson

Attempting to align AI capabilities and value structures by means of value elicitation from humans, such as through Reinforcement Learning from Human Feedback (RLHF), is a computational challenge that raises both psychological and philosophical questions. Adopting an evolutionary perspective on the emergence of value structures in humans and machine learning systems can offer a bridge between qualitative and quantitative aspects of alignment. Here, evolutionary dynamics are applied to a game-theoretic model of RLHF. This allows for formal reasoning about the process and capabilities that result from alignment training, even where quantitative benchmarks cannot be clearly defined. A simple parametrized game model of RLHF, subject to replicator dynamics, shows how the success of the training method is sensitive to bias in human judgments. Under ideal conditions, RHLF training leads to aligned behavior. If the choice pattern of the human judge is biased, the training instead incentivizes misalignment. This application shows that evolutionary analyses can contribute to improving the prospects for safety and support successful cooperation between humans and AI systems in deployment.

Zhengyang Lu, Po-Chun Chien, Nian-Ze Lee, Vijay Ganesh

We build the first machine-learning-based algorithm selection tool for hardware verification described in the Btor2 format. In addition to hardware verifiers, our tool also selects from a set of software verifiers to solve a given Btor2 instance, enabled by a Btor2-to-C translator. We propose two embeddings for a Btor2 instance, Bag of Keywords and Bit-Width Aggregation. Pairwise classifiers are applied for algorithm selection. Upon evaluation, our tool Btor2-Select solves 30.0% more instances and reduces PAR-2 by 50.2%, compared to the PDR implementation in the HWMCC'20 winner model checker AVR. Measured by the Shapley values, the software verifiers collectively contributed 27.2% to Btor2-Select's performance.

Xiangkui Lu, Jun Wu

Top-one recommendation with anonymous user behaviors, also known as session-based recommendation (SBR), faces challenges of top-one ranking and short anonymous sequences. To this end, we propose a novel objective that combines (1) a reciprocal rank loss to directly optimize the benchmark metric of top-one recommendation, with (2) a listwise contrastive loss to handle short sequences through listwise augmented consistency regularization. Empirical studies demonstrate that optimizing the proposed objective significantly improves the performance of existing SBR baselines.

Sijin Lu, Fangyuan Luo, Jun Wu

Causal Inference (CI) plays a crucial role in building unbiased recommender systems. However, most current CI-based debiasing methods only pay attention on either popularity bias or conformity bias. This paper presents a Disentangled Counterfactual Reasoning framework to alleviate dual biases in recommendation, so called DCR. Concretely, we consider the impact of both item popularity and user conformity during training, and separate their indirect effects by disentangling user and item embeddings into biased and unbiased components. In the inference stage, we perform counterfactual reasoning to simultaneously mitigate the indirect and direct effects of bias factors. Experimental results demonstrate the effectiveness of our DCR.

Hou-Wan Long, On-In Ho, Qi-Qiao He, Yain-Whar Si

Transfer learning enhances model performance in financial time series by leveraging data from related domains. The selection of appropriate source domains is crucial to avoid negative transfer. We propose using Gramian Angular Field (GAF) transformations to improve time series similarity functions for better domain alignment. Extensive experiments with DNN and LSTM models show that GAF-based similarity functions, specifically Coral (GAF) for DNN and CMD (GAF) for LSTM, significantly reduce prediction errors, demonstrating their effectiveness in complex financial environments.

Rally Lin, Christina Mo, Reyan Shariff, Darrick Zhang, Abdullah Alumar, Kaleb Kassaw, Leslie M. Collins, Boyla O. Mainsah

Brain-computer interfaces (BCIs) can provide a means of communication for individuals with severe neuromuscular diseases, the target end-users. While personalized BCI machine learning models are the current standard, models trained on data from other users could reduce BCI calibration time. We use a novel dataset with BCI users with and without amyotrophic lateral sclerosis (ALS) and a popular BCI deep learning model, EEGNet, to assess the impact of population domain data on transfer learning of a P300 speller task in the ALS cohort. Results show that training on source data from the non-ALS cohort was detrimental to transfer learning. In contrast, generic EEGNet models trained on source data from the ALS cohort performed comparably as user-specific models. Our findings highlight the need for more data from target end-users populations in publicly available BCI datasets.

Divya Lidder, Kathryn Morse, Bridget Sullivan, Wei Qian, Chenglin Miao, Mengdi Huai

Conformal prediction (CP) has gained prominence as a popular technique for uncertainty quantification in deep neural networks (DNNs), providing statistically rigorous uncertainty sets. However, existing CP methods fail to clarify the origins of predictive uncertainties. While neuron-level interpretability has been effective in revealing the internal mechanisms of DNNs, explaining CP at the neuron level remains unexplored. Nonetheless, generating neuron explanations for CP is challenging due to the discrete and non-differentiable characteristics of CP, and the labor-intensive process of semantic annotation. To address these limitations, this paper proposes a novel neuron explanation approach for CP by identifying neurons crucial for understanding predictive uncertainties and automatically generating semantic explanations. The effectiveness of the proposed method is validated through both qualitative and quantitative experiments.

Sarah Liaw

We propose a framework that uses renormalization group (RG) theory from statistical physics to analyze and optimize the hierarchical feature learning process in deep neural networks. Here, the layer-wise transformations in deep networks can be viewed as analogous to RG transformations, with each layer implementing a coarse-graining operation that extracts increasingly abstract features. We propose an approach to enforce scale invariance in neural networks, introduce scale-aware activation functions, and derive RG flow equations for network parameters. We show that our approach leads to fixed points corresponding to scale-invariant feature representations. Finally, we propose an RG-guided training procedure that converges to these fixed points while minimizing the loss function.

Lingfang Li, Procheta Sen

Recent advances in deep learning have expanded the application of large language models (LLMs) across fields such as medicine, finance, and education. Understanding the mechanisms underlying these models is essential to mitigate issues like hallucinations and bias. This study provides deep learning practitioners with insights into how specific training data points and internal structures influence model behaviour. Using influence functions and mechanistic interpretability, we will analyze the impact of data on model predictions across various tasks. Preliminary findings indicate that semantic search techniques, such as FAISS, enable efficient identification of influential training points in GPT-2 small. Future work will extend these methods to additional tasks and more complex models, with a focus on further elucidating LLM structures to improve interpretability.

Jonathan Li

Accurate forecasting of medication usage and ICD-9/10 code streams is critical for optimizing medical logistics, especially during periods of high demand, such as pandemics, disease outbreaks, wartime, or natural disasters. In this study, we develop a novel and robust forecasting framework using unsupervised learning techniques and Natural Language Processing (NLP) methods to build vector representations of daily ICD-9/10 codes and medication daily usage from Electronic Health Record (EHR) data. Multiple forecasting models, including Linear Drift Model, Vector Autoregression (VAR), Temporal Fusion Transformer (TFT), and Autoregressive Long Short-Term Memory (AR-LSTM) are trained, tested and evaluated. Finally multiple TFT and AR-LSTM models with different lookback horizon are trained and ensembled together to achieve better forecasting accuracy in near further (10 days). The AI framework is validated using MIMIC-IV ER and MIMIC-III datasets, resulting in the average forecasting error 5.2% at 5-th day and 18.1% at the 10-th day. The results demonstrate the ensemble model’s superior performance on near-future medication usage forecasting and ICD code progression, offering valuable insights for healthcare logistics and decision making. The framework also provides the mechanism to detect the model drift and finetune the model if necessary, which offers a robust tool for managing healthcare logistics under extreme and fluctuating conditions.

Taegyeom Lee, Ohyun Jo

This paper addresses contention window optimization for multi-access scenarios. Our investigation into state-of-the-art models revealed that a limited number of nodes dominate the communication channels. Such monopolization issues are critical in networks as they can lead to significant disruptions. To mitigate this monopolization problem, we propose an imitation learning-based backoff mechanism. The proposed model is a reinforcement learning-based contention window optimization method. It imitates the expert's policy to ensure fair policy convergence for the agent and includes opportunities for weight adjustment to boost performance. The proposed model shows a fairness improvement of approximately 20% to 41% across various scenarios.

Mitsuka Kiyohara

Increasing student populations and diverse course offerings have led to perceived inequities in U.S. high school course scheduling. Traditional integer programming (IP) methods for the High School Scheduling Problem (HSSP) fail to address these fairness concerns. This research introduces the Fair High School Scheduling Problem (FHSSP), an extension of the HSSP that incorporates student preferences and fairness principles from market design. We develop an IP model to generate course schedules that are both feasible and equitable. Tested on real course request data from a California high school, our model successfully produces schedules that ensure fairness without compromising feasibility. These results demonstrate the potential of our approach to enhance fairness in high school scheduling and its applicability to various real-world scheduling challenges. Additionally, this study highlights the feasibility of integrating human preferences and emotions into mathematical models, promoting more inclusive and balanced allocation systems.