no code implementations • NAACL (TextGraphs) 2021 • Parishad BehnamGhader, Hossein Zakerinia, Mahdieh Soleymani Baghshah
Pre-trained models like Bidirectional Encoder Representations from Transformers (BERT), have recently made a big leap forward in Natural Language Processing (NLP) tasks.
no code implementations • 9 Apr 2024 • Omid Ghahroodi, Marzia Nouri, Mohammad Vali Sanian, Alireza Sahebi, Doratossadat Dastgheib, Ehsaneddin Asgari, Mahdieh Soleymani Baghshah, Mohammad Hossein Rohban
Evaluating Large Language Models (LLMs) is challenging due to their generative nature, necessitating precise evaluation methodologies.
no code implementations • 27 Mar 2024 • Reza Abbasi, Mohammad Samiei, Mohammad Hossein Rohban, Mahdieh Soleymani Baghshah
Vision-language models, such as CLIP, have shown promising Out-of-Distribution (OoD) generalization under various types of distribution shifts.
no code implementations • 29 Feb 2024 • Fahimeh Hosseini Noohdani, Parsa Hosseini, Aryan Yazdan Parast, HamidReza Yaghoubi Araghi, Mahdieh Soleymani Baghshah
Based on our observations, models trained with ERM usually highly attend to either the causal components or the components having a high spurious correlation with the label (especially in datapoints on which models have a high confidence).
no code implementations • 8 Dec 2023 • Mahdi Ghaznavi, Hesam Asadollahzadeh, HamidReza Yaghoubi Araghi, Fahimeh Hosseini Noohdani, Mohammad Hossein Rohban, Mahdieh Soleymani Baghshah
In order to provide group robustness without such annotations, we propose a new method, called loss-based feature re-weighting (LFR), in which we infer a grouping of the data by evaluating an ERM-pre-trained model on a small left-out split of the training data.
no code implementations • 15 Nov 2023 • Reza Yarbakhsh, Mahdieh Soleymani Baghshah, Hamidreza Karimaghaie
Financial market forecasting remains a formidable challenge despite the surge in computational capabilities and machine learning advancements.
no code implementations • 11 Jan 2023 • Seyyed AmirHossein Ameli Kalkhoran, Mohammadamin Banayeeanzade, Mahdi Samiei, Mahdieh Soleymani Baghshah
The existing continual learning methods are mainly focused on fully-supervised scenarios and are still not able to take advantage of unlabeled data available in the environment.
1 code implementation • 29 Oct 2022 • Seyed Roozbeh Razavi Rohani, Saeed Hedayatian, Mahdieh Soleymani Baghshah
Inspired by recent progress in meta-RL, we introduce BIMRL, a novel multi-layer architecture along with a novel brain-inspired memory module that will help agents quickly adapt to new tasks within a few episodes.
no code implementations • 20 Sep 2022 • Faezeh Faez, Negin Hashemi Dijujin, Mahdieh Soleymani Baghshah, Hamid R. Rabiee
Deep learning-based graph generation approaches have remarkable capacities for graph data modeling, allowing them to solve a wide range of real-world problems.
no code implementations • 5 Aug 2022 • Sara Ghazanfari, Ali Rasteh, Seyed Abolfazl Motahari, Mahdieh Soleymani Baghshah
Most studies have shown that alternative splicing plays a significant role in human health and disease.
no code implementations • 8 Nov 2021 • Eduardo Conde-Sousa, João Vale, Ming Feng, Kele Xu, Yin Wang, Vincenzo Della Mea, David La Barbera, Ehsan Montahaei, Mahdieh Soleymani Baghshah, Andreas Turzynski, Jacob Gildenblat, Eldad Klaiman, Yiyu Hong, Guilherme Aresta, Teresa Araújo, Paulo Aguiar, Catarina Eloy, António Polónia
Breast cancer is the most common malignancy in women, being responsible for more than half a million deaths every year.
no code implementations • 7 Oct 2021 • Yassaman Ommi, Matin Yousefabadi, Faezeh Faez, Amirmojtaba Sabour, Mahdieh Soleymani Baghshah, Hamid R. Rabiee
With an increase in the number of applications where data is represented as graphs, the problem of graph generation has recently become a hot topic.
no code implementations • 29 Sep 2021 • Seyyede Fatemeh Seyyedsalehi, Mahdieh Soleymani Baghshah, Hamid R. Rabiee
Because of the complex relationship between the computational path of output variables in structured models, a feature can affect the value of output through other ones.
no code implementations • 10 May 2021 • Hassan Hafez-Kolahi, Behrad Moniri, Shohreh Kasaei, Mahdieh Soleymani Baghshah
For the upper bound, the optimization is further constrained to use $R$ bits from the training set, a setting which relates MER to information-theoretic bounds on the generalization gap in frequentist learning.
1 code implementation • 27 Feb 2021 • Mahsa Ghorbani, Anees Kazi, Mahdieh Soleymani Baghshah, Hamid R. Rabiee, Nassir Navab
This is accomplished by associating a graph-based neural network to each class, which is responsible for weighting the class samples and changing the importance of each sample for the classifier.
1 code implementation • 1 Feb 2021 • Sina Hajimiri, Aryo Lotfi, Mahdieh Soleymani Baghshah
We address this problem by proposing a framework capable of disentangling class-related and class-independent factors of variation in data.
no code implementations • 31 Dec 2020 • Faezeh Faez, Yassaman Ommi, Mahdieh Soleymani Baghshah, Hamid R. Rabiee
Deep generative models have achieved great success in areas such as image, speech, and natural language processing in the past few years.
no code implementations • ACL 2020 • Amirhossein Kazemnejad, Mohammadreza Salehi, Mahdieh Soleymani Baghshah
With its novel editor module, the model then paraphrases the input sequence by editing it using the extracted relations between the retrieved pair of sentences.
2 code implementations • 24 Aug 2019 • Ehsan Montahaei, Danial Alihosseini, Mahdieh Soleymani Baghshah
The proposed method has an iterative manner in which each new generator is definedbased on the last discriminator.
3 code implementations • NAACL 2019 • Ehsan Montahaei, Danial Alihosseini, Mahdieh Soleymani Baghshah
In this paper, we propose metrics to evaluate both the quality and diversity simultaneously by approximating the distance of the learned generative model and the real data distribution.
no code implementations • 21 Nov 2018 • Ehsan Montahaei, Mahsa Ghorbani, Mahdieh Soleymani Baghshah, Hamid R. Rabiee
However, this approach has not been extensively utilized for classifier training.
1 code implementation • 21 Nov 2018 • Mahsa Ghorbani, Mahdieh Soleymani Baghshah, Hamid R. Rabiee
The goal of graph embedding is to find a low dimensional representation of graph nodes that preserves the graph information.
no code implementations • 18 Jun 2016 • Amirhossein Akbarnejad, Mahdieh Soleymani Baghshah
Multi-label classifiers have to address many problems including: handling large-scale datasets with many instances and a large set of labels, compensating missing label assignments in the training set, considering correlations between labels, as well as exploiting unlabeled data to improve prediction performance.
no code implementations • 29 May 2016 • Seyed Mohsen Shojaee, Mahdieh Soleymani Baghshah
We seek a linear transformation on signatures to map them onto the visual features, such that the mapped signatures of the seen classes are close to labeled samples of the corresponding classes and unlabeled data are also close to the mapped signatures of one of the unseen classes.
no code implementations • 20 Jan 2016 • Mitra Montazeri, Mahdieh Soleymani Baghshah, Aliakbar Niknafs
Hyper-heuristic is a new heuristic approach which can search the solution space effectively by applying local searches appropriately.
no code implementations • 15 Dec 2015 • Mitra Montazeri, Mahdieh Soleymani Baghshah, Ahmad Enhesari
In our study, new method is proposed for identify efficient features of lung cancer.
no code implementations • 12 Dec 2015 • Mehrdad Ghadiri, Amin Aghaee, Mahdieh Soleymani Baghshah
k-medoids algorithm is a partitional, centroid-based clustering algorithm which uses pairwise distances of data points and tries to directly decompose the dataset with $n$ points into a set of $k$ disjoint clusters.