Md. Kowsher
YOU?
Author Swipe
View article: SliceFine: The Universal Winning-Slice Hypothesis for Pretrained Networks
SliceFine: The Universal Winning-Slice Hypothesis for Pretrained Networks Open
This paper presents a theoretical framework explaining why fine tuning small, randomly selected subnetworks (slices) within pre trained models can be sufficient for downstream adaptation. We prove that pretrained networks exhibit a univers…
View article: Explainable Detection of Implicit Influential Patterns in Conversations via Data Augmentation
Explainable Detection of Implicit Influential Patterns in Conversations via Data Augmentation Open
In the era of digitalization, as individuals increasingly rely on digital platforms for communication and news consumption, various actors employ linguistic strategies to influence public perception. While models have become proficient at …
View article: Predicting Through Generation: Why Generation Is Better for Prediction
Predicting Through Generation: Why Generation Is Better for Prediction Open
This paper argues that generating output tokens is more effective than using pooled representations for prediction tasks because token-level generation retains more mutual information. Since LLMs are trained on massive text corpora using n…
View article: TituLLMs: A Family of Bangla LLMs with Comprehensive Benchmarking
TituLLMs: A Family of Bangla LLMs with Comprehensive Benchmarking Open
In this paper, we present TituLLMs, the first large pretrained Bangla LLMs, available in 1b and 3b parameter sizes. Due to computational constraints during both training and inference, we focused on smaller models. To train TituLLMs, we co…
View article: User Profile with Large Language Models: Construction, Updating, and Benchmarking
User Profile with Large Language Models: Construction, Updating, and Benchmarking Open
User profile modeling plays a key role in personalized systems, as it requires building accurate profiles and updating them with new information. In this paper, we present two high-quality open-source user profile datasets: one for profile…
View article: BnTTS: Few-Shot Speaker Adaptation in Low-Resource Setting
BnTTS: Few-Shot Speaker Adaptation in Low-Resource Setting Open
This paper introduces BnTTS (Bangla Text-To-Speech), the first framework for Bangla speaker adaptation-based TTS, designed to bridge the gap in Bangla speech synthesis using minimal training data. Building upon the XTTS architecture, our a…
View article: Parameter-efficient fine-tuning of large language models using semantic knowledge tuning
Parameter-efficient fine-tuning of large language models using semantic knowledge tuning Open
Large Language Models (LLMs) are gaining significant popularity in recent years for specialized tasks using prompts due to their low computational cost. Standard methods like prefix tuning utilize special, modifiable tokens that lack seman…
View article: Does Self-Attention Need Separate Weights in Transformers?
Does Self-Attention Need Separate Weights in Transformers? Open
The success of self-attention lies in its ability to capture long-range dependencies and enhance context understanding, but it is limited by its computational complexity and challenges in handling sequential data with inherent directionali…
View article: LLM-Mixer: Multiscale Mixing in LLMs for Time Series Forecasting
LLM-Mixer: Multiscale Mixing in LLMs for Time Series Forecasting Open
Time series forecasting remains a challenging task, particularly in the context of complex multiscale temporal patterns. This study presents LLM-Mixer, a framework that improves forecasting accuracy through the combination of multiscale ti…
View article: RoCoFT: Efficient Finetuning of Large Language Models with Row-Column Updates
RoCoFT: Efficient Finetuning of Large Language Models with Row-Column Updates Open
We propose RoCoFT, a parameter-efficient fine-tuning method for large-scale language models (LMs) based on updating only a few rows and columns of the weight matrices in transformers. Through extensive experiments with medium-size LMs like…
View article: Parameter-Efficient Fine-Tuning of Large Language Models using Semantic Knowledge Tuning
Parameter-Efficient Fine-Tuning of Large Language Models using Semantic Knowledge Tuning Open
Large Language Models (LLMs) are gaining significant popularity in recent years for specialized tasks using prompts due to their low computational cost. Standard methods like prefix tuning utilize special, modifiable tokens that lack seman…
View article: Propulsion: Steering LLM with Tiny Fine-Tuning
Propulsion: Steering LLM with Tiny Fine-Tuning Open
The rapid advancements in Large Language Models (LLMs) have revolutionized natural language processing (NLP) and related fields. However, fine-tuning these models for specific tasks remains computationally expensive and risks degrading pre…
View article: Token Trails: Navigating Contextual Depths in Conversational AI with ChatLLM
Token Trails: Navigating Contextual Depths in Conversational AI with ChatLLM Open
Conversational modeling using Large Language Models (LLMs) requires a nuanced understanding of context to generate coherent and contextually relevant responses. In this paper, we present Token Trails, a novel approach that leverages token-…
View article: Changes by Butterflies: Farsighted Forecasting with Group Reservoir Transformer
Changes by Butterflies: Farsighted Forecasting with Group Reservoir Transformer Open
In Chaos, a minor divergence between two initial conditions exhibits exponential amplification over time, leading to far-away outcomes, known as the butterfly effect. Thus, the distant future is full of uncertainty and hard to forecast. We…
View article: L-TUNING: Synchronized Label Tuning for Prompt and Prefix in LLMs
L-TUNING: Synchronized Label Tuning for Prompt and Prefix in LLMs Open
Efficiently fine-tuning Large Language Models (LLMs) for specific tasks presents a considerable challenge in natural language processing. Traditional methods, like prompt or prefix tuning, typically rely on arbitrary tokens for training, l…
View article: Short-Term Rainfall Prediction Using Supervised Machine Learning
Short-Term Rainfall Prediction Using Supervised Machine Learning Open
Floods and rain significantly impact the economy of many agricultural countries in the world. Early prediction of rain and floods can dramatically help prevent natural disaster damage. This paper presents a machine learning and data-driven…
View article: BENGALI INFORMATION RETRIEVAL SYSTEM (BIRS)
BENGALI INFORMATION RETRIEVAL SYSTEM (BIRS) Open
Information Retrieval System is an effective process that helps a user to trace relevant information by Natural Language Processing (NLP). In this research paper, we have presented present an algorithmic Information Retrieval System(BIRS) …
View article: Contrastive Learning for Universal Zero-Shot NLI with Cross-Lingual Sentence Embeddings
Contrastive Learning for Universal Zero-Shot NLI with Cross-Lingual Sentence Embeddings Open
Natural Language Inference (NLI) is a crucial task in natural language processing, involving the classification of sentence pairs into entailment, contradiction, or neutral categories.This paper introduces a novel approach to achieve unive…
View article: A Comprehensive Review on Big Data for Industries: Challenges and Opportunities
A Comprehensive Review on Big Data for Industries: Challenges and Opportunities Open
Technological advancements in large industries like power, minerals, and manufacturing are generating massive data every second. Big data techniques have opened up numerous opportunities to utilize massive datasets in several effective way…
View article: Impact Learning: A Learning Method from Features Impact and Competition
Impact Learning: A Learning Method from Features Impact and Competition Open
Machine learning is the study of computer algorithms that can automatically improve based on data and experience. Machine learning algorithms build a model from sample data, called training data, to make predictions or judgments without be…
View article: Transfer Learning for Sentiment Analysis Using BERT Based Supervised Fine-Tuning
Transfer Learning for Sentiment Analysis Using BERT Based Supervised Fine-Tuning Open
The growth of the Internet has expanded the amount of data expressed by users across multiple platforms. The availability of these different worldviews and individuals’ emotions empowers sentiment analysis. However, sentiment analysis beco…
View article: A review on job scheduling technique in cloud computing and priority rule based intelligent framework
A review on job scheduling technique in cloud computing and priority rule based intelligent framework Open
In recent years, the concept of cloud computing has been gaining traction to provide dynamically increasing access to shared computing resources (software and hardware) via the internet. It’s not secret that cloud computing’s ability to su…
View article: An Enhanced Neural Word Embedding Model for Transfer Learning
An Enhanced Neural Word Embedding Model for Transfer Learning Open
Due to the expansion of data generation, more and more natural language processing (NLP) tasks are needing to be solved. For this, word representation plays a vital role. Computation-based word embedding in various high languages is very u…
View article: CARAN: A Context-Aware Recency-Based Attention Network for Point-of-Interest Recommendation
CARAN: A Context-Aware Recency-Based Attention Network for Point-of-Interest Recommendation Open
Point-of-interest (POI) recommendation system that tries to anticipate user’s next visiting location has attracted a plentiful research interest due to its ability in generating personalized suggestions. Since user’s historical check-ins a…