VoyageVista

Inspiration

The inspiration for our project stemmed from recognizing the common challenges faced by travelers in terms of remembering flight details, navigating the intricate planning process, and dealing with the stress associated with these aspects. We were motivated to leverage the capabilities of Language Model (LLM) technology to create an innovative solution that acts as a comprehensive travel assistant, enhancing user experience and facilitating smoother journeys.

What it does

Our travel assistant chatbot utilizes Language Models to offer a range of functionalities:

Enhanced LLM Interaction: Collaborating with the LangChain framework, our chatbot provides dynamic and engaging user interactions, ensuring a personalized and user-friendly dialogue experience.
Flight Information Retrieval: Leveraging vector similarity search, our system efficiently retrieves and presents user flight ticket information, eliminating the need for users to manually search for these details.
Travel Suggestion: Provide the travel suggestion based on the attractions that available in Journify.
Aspect-Based Sentiment Analysis: The chatbot employs Aspect-Based Sentiment Analysis to gain insights into user perspectives based on their feedback. This not only helps in understanding user sentiment but also facilitates continuous improvement of the travel assistant.

How we built it

We built our travel assistant chatbot by integrating various technologies and methodologies:

For the enhanced LLM, we collaborated with the LangChain framework, implementing it to enable dynamic and personalized user interactions. Vector similarity search was employed to efficiently retrieve user flight ticket information, enhancing the speed and accuracy of information retrieval. Aspect-Based Sentiment Analysis was integrated into the chatbot to analyze user feedback and sentiments related to different aspects of their travel experience.

Challenges we ran into

Integration Complexity: Implementing modular LLM backends on Databricks poses integration challenges, requiring seamless coordination between different components and third-party frameworks. Resource Management: Efficiently managing resources on Databricks for training and deploying large language models may present scalability and performance hurdles. Versioning and Compatibility: Ensuring compatibility between the language models, frameworks (e.g., LangChain), and Databricks versions is critical to prevent conflicts and maintain stability. Data Handling: Addressing data preprocessing challenges and managing large datasets within the Databricks environment can be complex, impacting the training and testing phases. Cost Optimization: Optimizing costs associated with Databricks usage, particularly for resource-intensive tasks like training LLMs, requires careful consideration and monitoring.

Accomplishments that we're proud of

We take pride in achieving the following milestones:

Successful integration of the LangChain framework for dynamic and engaging LLM interactions. Efficient implementation of vector similarity search for quick retrieval of flight information. Successful deployment of Aspect-Based Sentiment Analysis, providing valuable insights into user sentiments for continuous improvement.

What we learned

Seamless Tool Integration Mastery: Gain a deeper understanding of LangChain features like prompts, chains, and agents, unlocking enhanced capabilities in developing dynamic conversational experiences. Langchain Agent Synergy: A key takeaway is the adept combination of the Langchain agent with OpenAI for handling CSV data. This synergy offers a deeper understanding of how to leverage different technologies for enhanced language processing and data analysis.

What's next for Aspect Based Sentiment Analysis

The future roadmap for Aspect-Based Sentiment Analysis includes:

Refinement of Models: Continuous refinement of sentiment analysis models to improve accuracy and adaptability to diverse user inputs. Expansion of Analyzed Aspects: Enhancing the system to analyze sentiment across a broader range of aspects, providing more detailed insights into user feedback. Integration with Continuous Learning: Implementing a continuous learning system that adapts and evolves based on user interactions and feedback, ensuring ongoing improvement in user satisfaction and experience.