IEEE SBLTW 2025で話す内容を準備する

I'm researching Ambient Assisted Living (AAL) to reduce the burden of caregiving. My work supports the independent living of elderly people by using sensors. Specifically, I am trying to recognize their activities through Sound Event Detection using microphones. Since the data obtained from these sensors is highly private, I aim to apply Federated Learning.

簡単な表現

I want to make life easier for people who take care of others.

So, I am studying AAL (Ambient Assisted Living).

This helps older people live safely on their own using sensors.

I use microphones to listen for sounds and understand what they are doing. This is called Sound Event Detection.

The sensor data is very private.

So, I want to use a special method called Federated Learning to protect their privacy.

burden か workload

連合学習のちょっとした説明

With this method, we don't collect raw data. Instead, we only gather the parameters, which are the learning results.

This way, we can train the model while protecting privacy

TRONの説明

So, TRON is Real-Time Operating System.

This means it works very fast and at the exact right time.

It's used in many things, like car engines, printers, and cameras.

内容

I am researching AAL to reduce workload of caregiving.

My works support independent living of elderly pepole using sensors.

Specificality, I am tring recognize theire activities by Sound Event Detection using micropone.

This way has one probrem, which is privacy of data.

That why, I aim to apply Federated Learning.

It do not gather the data at one point. Instead, gather the parametera.

So, we can enshure the privacy and learn model by local data.