Presented by: Hang Ruan

Date: Thursday, October 10, 2024
Time: 9:00 am
Place: Zoom, link below.

Abstract: Sentiment analysis is a critical task in Natural Language Processing (NLP) that helps decode the emotions and opinions embedded in text. With applications spanning from market research and social media monitoring to political analysis and customer feedback evaluation, sentiment analysis provides invaluable insights into public opinion and consumer behavior. This report studies the evolution of sentiment analysis models, focusing on the advancements made by deep learning techniques such as Long Short-Term Memory (LSTM), Convolutional Neural Networks (CNNs), and transformer-based models like Bidirectional Encoder Representations from Transformers (BERT) and Generative Pre-trained Transformer (GPT). These models have set new benchmarks for accuracy, efficiency, and versatility. Additionally, this work explores Mamba, a recent State Space Model (SSM) designed to overcome the computational challenges of transformers in handling long sequences and demonstrates state-of-the-art performance on language modeling tasks that is comparable to transformers twice its size. This study critically examines the strengths and limitations of these models, comparing their performance on sentiment analysis datasets to provide a comprehensive understanding of their applicability and efficacy in various contexts.
 

Topic: Zoom Meeting

Time: Oct 10, 2024 09:00 AM Pacific Time (US and Canada)

Join Zoom Meeting

https://uvic.zoom.us/j/81185533518?pwd=m1IBwarDY2MbZXco1wgZhya0RMS7xb.1

Meeting ID: 811 8553 3518

Password: 090223

One tap mobile

+16475580588,,81185533518# Canada

+17789072071,,81185533518# Canada

Dial by your location

        +1 647 558 0588 Canada

        +1 778 907 2071 Canada

Meeting ID: 811 8553 3518

Find your local number: https://uvic.zoom.us/u/kciJywqhDy

Presented by: Jinke Jiang

Date: Friday, October 11, 2024
Time: 2:00 pm
Place: Zoom, link below.

Abstract: As a result of population aging, the number of visually impaired people is growing. Unfortunately, there is limited accessibility measures to help improve the quality of life of these people. The recent technological development in Artificial Intelligence, especially Large Language Models (LLMs), should offer effective and efficient solutions. Recognizing the limitation of existing products, we design and implement a user-friendly and privacy-safe voice assistant for visually impaired people. Using LLMs and Vision Language Models, the assistant can recognize and identify objects through low-latency speech-to-speech interactions. The assistant can be deployed on offline edge computing devices with camera/microphone/speaker, with easily extendable functionalities. In this report, we present the design, adopted technologies, and adjustment that we applied to arrive at the final implementation
 

Topic: Zoom Meeting

Time: Oct 11, 2024 02:00 PM Vancouver

Join Zoom Meeting

https://uvic.zoom.us/j/89457089273?pwd=a1BbHMJQalkUwdbtakEcaa3BS8yIpp.1

Meeting ID: 894 5708 9273

Password: 195461

One tap mobile

+17789072071,,89457089273# Canada

+16475580588,,89457089273# Canada

Dial by your location

        +1 778 907 2071 Canada

        +1 647 558 0588 Canada

Meeting ID: 894 5708 9273

Presented by: Thet Naung San

Date: Monday, October 14, 2024
Time: 8:30 am
Place: Zoom, link below.

Abstract: The increasing demand for low-latency, high-reliability communications in 5G New Radio (NR) is driving the need for optimized transmission technologies, especially for Ultra-Reliable Low-Latency Communication (URLLC) applications. This project introduces a Deep Q-Network (DQN) agent that intelligently adjusts the parameters of modulation and coding scheme (MCS) and numerology adaptively in real time to improve physical downlink shared channel (PDSCH) transmissions. The system simulates frequency-selective TDL-C fading channels with a Hybrid Automatic Repeat Request (HARQ) mechanism and optimizes based on packet size, signal-to-noise ratio (SNR), latency and number of HARQ retransmissions. The results show improved transmission efficiency and reliability, providing better performance for mission-critical 5G URLLC applications.

Thet Naung San is inviting you to a scheduled Zoom meeting.

Topic: ECE Graduate Seminar - Thet Naung San

Time: Oct 14, 2024 08:30 AM Pacific Time (US and Canada)

Join Zoom Meeting

https://uvic.zoom.us/j/84250350747?pwd=5ybirUkAQL68j2wo5Xt9VoywWLNwRn.1

Meeting ID: 842 5035 0747

Password: 112477

One tap mobile

+17789072071,,84250350747#,,,,0#,,112477# Canada

+16475580588,,84250350747#,,,,0#,,112477# Canada

Dial by your location

        +1 778 907 2071 Canada

        +1 647 558 0588 Canada

Meeting ID: 842 5035 0747

Password: 112477

Find your local number: https://uvic.zoom.us/u/kc1u3nndfN

Presented by: Jiachao Deng

Date: Thursday, October 17, 2024
Time: 2:30 pm
Place: Zoom, link below.

Abstract:

C++ source code analysis is conducted at scale. A framework is proposed for analyzing the C++ codebase of operating systems that employ the dnf package manager, such as Fedora Linux and Red Hat Enterprise Linux. The framework can run an arbitrary static analysis tool over software packages that contain C++ code from compatible operating systems. In order to evaluate the effectiveness of the framework and to better understand how the C++ language is used in practice, a C++ analysis tool is developed to study library usage with a fine level of granularity, considering instances of uses of types, type aliases, member/non-member functions, variables, and enumerators.

The framework, combined with the C++ library usage analysis tool, is used to analyze 2379 software packages from the codebase of Fedora Linux 37. A total of 398065762 lines of C++ code is investigated, a scale not achieved by previous C++ research. Based on the Clang compiler front-end libraries, our library usage analysis tool addresses C++ parsing accuracy issues found in many other studies. Consequently, the tool extracts a reliable collection of library usage instances from C++ software.

Numerous observations are made regarding various aspects of library usage that can facilitate improved teaching of C++, aid in the refinement of C++ libraries, and help guide the future evolution of the C++ standard. For example, our analysis reveals that C++ programmers rarely use some C++ standard library algorithms designed for specialized purposes or combined operations. These algorithms often appear in less than 1% of all C++ software packages investigated. The low adoption rate suggests possible improvements by deprecating such algorithms to simplify the standard library interface. Such observations summarize current trends in C++ library usage and provide recommendations for improving the C++ language and its libraries.

Jiachao Deng is inviting you to a scheduled Zoom meeting.

Topic: Jiachao Deng, ECE Graduate Seminar, An Analysis of Library Usage in the C++ Code Base of Fedora Linux 37

Time: Oct 17, 2024 02:30 PM Pacific Time (US and Canada)

Join Zoom Meeting

https://uvic.zoom.us/j/84043181194?pwd=jL72Zz49hJD3WOhcgMbpiFv3Gi0Jjh.1

Meeting ID: 840 4318 1194

Password: 277970

One tap mobile

+16475580588,,84043181194# Canada

+17789072071,,84043181194# Canada

Dial by your location

       +1 647 558 0588 Canada

       +1 778 907 2071 Canada

Meeting ID: 840 4318 1194

Find your local number: https://uvic.zoom.us/u/krLfnxN4d