ACCT 331 WEEK 11

Class Schedule:
- Location: Schreiber #302
- Days: Tuesday & Thursday
- Time: 1:00 - 2:15
- Course Code: 10304

Week 1: Introduction to Artificial Intelligence
- Overview of AI concepts and applications.
Important Dates:
- Exam #2: Week 12 on 11/13/25
- Final Exam: Weeks 12-14 on 12/12/00/25

Build Your Own Small Language Model
- Duration: 6 hours
- Fundamentals of language models and machine learning basics.
Train a Small Language Model (Challenge Lab)
- Duration: 1 hour 30 minutes
- Focus on developing tools and data preparation.
Represent Your Language Data
- Duration: 4 hours
- Preparation of text data for language modeling.
Design And Train Neural Networks
- Duration: 4 hours
- In-depth focus on the training process for machine learning models.
Discover The Transformer Architecture
- Duration: 4 hours
- Mechanisms and applications of the transformer architecture.
AI Research Foundations
- Course covering foundational knowledge for AI research.

JPMorgan Chase's AI Investment
- Jamie Dimon states a $2 billion investment in AI has paid for itself.
- Investment leads to operational savings across various business lines.
KPMG Examining AI Usage
- AI tool usage will impact annual performance reviews for employees.
- All employees assessed on how they integrate AI tools into their work.
Rising Tech Investments
- Chart showing an increase in annualized capital expenditures across big tech companies like Meta and Microsoft.
- Notable growth in data centers construction spending as well.
Usage Growth for ChatGPT
- ChatGPT projected growth in user metrics, indicating significant adoption in various sectors.

Many industries, such as financial services and consulting, are seeing transformative changes in operations and employee roles due to AI.
Major firms are increasing pressure on staff to incorporate AI tools for efficiency and performance growth.

Tuesday: Introduction to deep learning.
Thursday: Foundations of deep learning.
- Data pipeline and self-supervised learning techniques.
- Introduction to Large Language Models (LLMs).

Tuesday: Further exploration of LLMs and their first principles.
Thursday: Focus on Generative AI (GenAI), transformers, and evaluation metrics.

Language Models (LMs): Use AI to predict subsequent words in a sentence based on contextual understanding.
- Example: Asking a question to a search engine results in predictive text responses.
Self-Supervised Learning (SSL): A technique for training without labeled data, where algorithms learn from automatic data modifications to create labels dynamically.

Understanding the nature and application of language models is essential for effective AI systems.
Capable of profoundly impacting fields like natural language processing (NLP), including applications in marketing, customer service, and knowledge management.

Cost Reduction: Automates repetitive tasks leading to significant savings (70-85%).
Revenue Growth: New product capabilities and personalized experiences increase conversion rates by 15-30%.
Efficiency: Accelerates knowledge work and decision-making processes, resulting in 40-60% time savings.

Architecture of LLMs relies heavily on transformers, which enhance understanding and processing of language through mechanisms like attention.
Attention Mechanism: It allows the model to weigh the importance of different words based on their relevance.

Involves several stages including tokenization, embedding, self-attention, multi-head attention, and residual connections to achieve rich context-aware representations.

Customer Service Automation: Utilized LLMs to improve response accuracy significantly.
Content Generation: Hybrid evaluation frameworks adopted for brand consistency in AI-generated content.
Knowledge Management: Enhancements in internal knowledge retrieval systems boosted performance metrics.

Ethical implications, biases in AI outputs, and the need for model interpretability are ongoing challenges.
Requirements for continual adaptation and evaluation of AI systems to enhance understanding and mitigate risks.

The course emphasizes the mathematical, structural, and practical implications of Applied Artificial Intelligence from a holistic viewpoint, preparing students for a future where AI is intricately tied to business success.