Question | Would it be a good idea to share an embedding layer between inputs and labels?

Let's say in the movielens dataset I have another column where each row is the list of the last 100 movies clicked by that user.

Would it make sense to encode the movie_watched_id in the candidate tower using the embedding layer used in the query tower to embed last_100_movies_clicked_ids?

Like, having this user model:

class UserModel(tf.keras.Model):

  def __init__(self):
    super().__init__()

    self.user_embedding = tf.keras.Sequential([
        tf.keras.layers.experimental.preprocessing.StringLookup(
            vocabulary=unique_user_ids, mask_token=None),
        tf.keras.layers.Embedding(len(unique_user_ids) + 1, 32),
    ])

    self.last_100_movies_clicked_ids_embedding = tf.keras.Sequential([
        tf.keras.layers.experimental.preprocessing.TextVectorizer(
            vocabulary=unique_last_100_movies_clicked_ids, mask_token=None),
        tf.keras.layers.Embedding(len(unique_last_100_movies_clicked_ids) + 1, 32),
    ])

  def call(self, inputs):
    # Take the input dictionary, pass it through each input layer,
    # and concatenate the result.
    return tf.concat([self.user_embedding(inputs["user_id"]), 
   					  self.last_100_movies_clicked_ids_embedding(inputs["last_100_movies_clicked_ids"])
        			])

We define the movie model as:

class MovieModel(tf.keras.Model):

  def __init__(self, user_model):
    super().__init__()

    max_tokens = 10_000

    self.movie_watched_id_embedding = user_model.last_100_movies_clicked_ids_embedding

  def call(self):
    return self.movie_watched_id_embedding(inputs['movie_watched_id'])