Foodbot: Empowering Healthy Eating through Goal-oriented Chatbot Intervention
With increasing attention on healthy eating habits, there is still a lack of tools to motivate, personalise and provide timely support to those who need it in the general population. Personalized guidance considers preferences and dietary restrictions for long-term adherence as well as helping to prevent diet-related diseases. With more personalized and timely support, it can motivate individuals to sustain healthy eating habits.
skillER: Effortless Skill Mention Extraction from Job Ads
Identifying skills mentioned in a job advertisement is crucial for both job seekers and employers. For job seekers, it helps align qualifications with the position, increasing their chances of being shortlisted. Employers benefit by attracting suitable candidates, streamlining the hiring process, and making better hiring decisions. It optimizes recruitment practices and ensures a good fit between candidates and job requirements.
Facial Weight Progression and Regression with Deep Learning
With the rise of digital platforms and online content consumption, there is a growing demand for tools that facilitate the customisation of engaging digital content. Traditional approaches may require significant amounts of labeled image data and may not generate realistic or high-quality transformations. This introduces an opportunity to introduce a framework that overcomes these limitations and provides a solution for seamless and realistic facial weight adjustments with deep learning.
Generative Pre-training Based Cooking Recipe Generation and Evaluation System
Advancements in AI have fueled a growing interest in the automatic generation of cooking recipes. With the abundance of online cooking recipes, a cutting-edge application called CulinAI—an intelligent system for recipe generation and evaluation has been developed that leverages AI to create innovative cooking recipes and instructions. CulinAI offers users two powerful modes. Firstly, instruction generation and secondly, ingredient generation from recipe.
Food Image Recognition for Asian Food
With greater emphasis on healthy living and healthier food choices, there is a need to track calorie and nutritional intake to better manage balanced diets and improve healthy lifestyle. Traditional and manual methods of identifying food types can be time-consuming and error prone. This technology aims to overcome these challenges by utilising advanced deep learning techniques for efficient and reliable food image recognition.
Occupancy Aware, Camera-Based Smart Lighting System
In workspaces and commercial buildings, lighting represents a significant portion of energy consumption after HVAC systems. However, current lighting solutions often lack efficiency and fail to adapt to occupancy levels and varying lighting needs. There is a need for an intelligent lighting system that can reduce energy consumption while maintaining occupant comfort by automatically adjusting lighting intensity based on occupancy and natural light levels.
Document-based Question Answering System
Finding precise answers to questions within given articles is a challenge that existing systems struggle to address. There is a need for an AI-driven question answering technology that can automatically extract short spans of text as precise answers from articles. We have developed a cutting-edge AI algorithm that powers a document-based question answering application. Our technology utilizes deep learning techniques to automatically extract answers from articles.
Quantum Machine Learning for Credit Scoring
Credit rating for small and medium-sized enterprises (SMEs) presents a challenge compared to larger, well-analyzed companies. There is a need for a more accurate and efficient credit rating system that can leverage the potential of big data and overcome the limitations of traditional approaches. We introduce QML models, utilize quantum computers for credit data analysis in Singaporean companies.
Consensus in distributed ledger technology is often limited by the blockchain trilemma, which encompasses data size, consensus speed, and network security. We present a leaderless quantum consensus algorithm that utilizes quantum simulators and real quantum computers for a simplified consensus scenario. We have developed a technology that harnesses the power of quantum computers and quantum communication channels for achieving consensus.
W8-Scope: Fine-Grained Monitoring of Weight Stack-based Exercises
Fine-grained and personalized monitoring of gym exercises is essential for users to track their progress, improve their routines, and prevent injuries. Existing approaches using wearable sensors or video sensing have limitations in terms of user engagement, privacy concerns, and accuracy in multi-user settings. There is a need for a cost-effective and unobtrusive technology that provides accurate insights into multiple aspects of gym exercise behavior.