What is the Value Add?
Value Add Overview
- Immediate Value
- Ongoing Value
- Long Term Value
1. Immediate Value
- Turns manual processes on the order of weeks into days
- Literal Annotation Interface
Turns manual multi-week processes into days
In comparison to blocking manual processes that take on the order of weeks to complete, with Diffgram your data science team can start working on algorithms as soon as a "Seed" set is completed. This turns a multi-week startup process into days . This benefit is for every single dataset!
Upon signing up for Diffgram, within minutes an admin user can create Task Templates and Datasets relevant to your business use case. Once this is created, the system will mange the entire data ingestion and task creation process as new data becomes available.
Create complex pipelines by simply daisy chaining Datasets and Task Templates. Diffgram automatically builds the complex non-linear graph for you. Think of this like setting up a Database but leaving it up to Postgres's Query Planner on how to actually retrieve the data.
Data Annotation Software has changed a lot. The breadth of interfaces and depth of assist options is exploding! Diffgram now has Integrations with ScaleAI, Datasaur, and Labelbox.
A data science team's one time effort to integrate with Diffgram now unlocks a growing ecosystem of tooling. For example a team monitoring crops can pull data from AWS sensors, get initial annotations through ScaleAI, have experts review inside Diffgram, then train on GCP - all from one place.
With Diffgram as the 'backend' for the data the combined solution available covers a huge range of use cases.
2. Ongoing Value
All of the Immediate Value and:
- Data Iteration and Reuse
- Changeable Process
- In Depth Reporting
Everything Data, including Iteration, Integration, and Reuse
- Upgrade data at different stages in the pipeline. Such as one stage doing a spatial bounding box and another adding attributes.
- The more models, more dataset variations, more users, more labels, more changes etc, all result in more benefits from the system.
- Powerful integrations for continuous integration and retraining
Compare this to only being able to linearly upgrade data and losing the ability to "time travel" across the dataset without extensive custom work.
- Redirect work to new tasks at any time, while leaving the old state "ready to resume".
- Easily update in progress templates, and identify data from already completed work that needs new templates.
Without a system this is hard to do for a single set. And virtually intractable for multiple sets. Diffgram allows the admin users to do this with no engineering intervention. This usage of the system is vastly different and more useful then a straight forward linear triggers at the set level.
The cycle of creating datasets is 1. Create set. 2. Train Algorithm 3. Create new set with knowledge gained from 2. With Diffgram it's easy to change these processes on the fly.
In Depth Reporting
Gain insight into project status, dataset make up and more.
- Strategic level, multiple project and multiple task overviews
- Deep drill downs into specific examples.
Diffgram has the most powerful reporting capabilities. The Abstract Report Generator is fully customizable by your project admins with no programming required.
3. Long Term Value
All of Immediate and Ongoing Value And:
- Query data with relationships maintained through time
- UI Driven feature set makes ongoing maintenance easy
- Our vision of making AI accessible and practical
- Query data be automatic system defined relationships
- Construct data relationships
UI Driven feature set
- Add new data sources without extra engineering
- Smooth operations for adding, removing users, organizing and archiving work.
Our vision of making AI accessible and practical
We believe Artificial Intelligence (AI) should be in every system because it automates knowledge tasks — leading to more creative work and multiplying the effectiveness of rare knowledge. Without supervision, AI Deep Learning systems don’t work. We create software for AI supervision.
Improving Training Data simultaneously helps improve Health care by extending doctors reach, improves Agriculture harvesting to feed more people, and extends top sports coaches knowledge to aspiring players. The faster we can improve and adopt Training Data the faster the adoption of these applications.
Updated over 2 years ago